Mathematical Reasoning Benchmarks

Follow this topic Ask about this topic

Momentum

13 papers in the last four weeks, up 117% on the four weeks before. 0.2% of all new papers.

Jul 6Week of Sep 21

Latest papers 101

Oct 1, 2026cs.AI

Improving Math Reasoning through Value-guided Informative Search

Reinforcement learning with verifiable rewards (RLVR) has substantially improved the mathematical reasoning capabilities of large language models. Recent work introduces search into RLVR rollouts to increase trajectory diversity, but diversity alone does not ensure that the search-induced rollout policy improves upon the current policy. To address this gap, we propose APIVIS, a training-time framework that adapts finite-budget Gumbel search to chunk-level mathematical reasoning. APIVIS combines direct and searched responses within each rollout group, allowing improvements found by search to produce informative relative rewards. It further applies selective supervision to search-improved tokens, preserving a learning signal when uniform group rewards render GRPO ineffective. We show that exact value-guided selection improves the expected verifier reward at each searched state and that this guarantee extends to the complete rollout policy, with a corresponding approximate guarantee under bounded value-estimation error. Experiments on widely recognized mathematical reasoning benchmarks and different model scales demonstrate substantial improvements over competitive search-based methods, validating the effectiveness of APIVIS.

Shaohuai Liu, Yuning Wu, Haoran Liu +3

Amazon · Department of Electrical and Computer Engineering, Texas A&M University

Mathematical Reasoning Benchmarks Reinforcement Learning With Verifiable Reward Verifiable Rewards

Sep 30, 2026cs.CL

Reason in Style: Discovering and Controlling Style in Language Models

Language models learn content and style jointly, making stylistic variation in their outputs difficult to identify and control. We study whether recurring styles in model responses can be discovered without supervision and explicitly controlled. We design an algorithm that learns to separate representations of content and style from language models' outputs and validate its effectiveness on math questions in a controlled setting. By applying this method to over 100K verified traces from nine distinct teacher models, we discover six recurring yet imbalanced styles. We then fine-tune smaller student models to follow these styles when explicitly conditioned on them, using importance weighting to balance the contribution of the styles represented in the corpus. This approach improves Pass@

k

over standard fine-tuning on the same data across six math reasoning benchmarks, demonstrating that we can diversify the style of answers effectively. We confirm that this also results in strong correspondence between requested and realized styles. We find that style affects correctness: the probability of solving a problem depends on the style we condition on, and different problems benefit from different styles. In summary, our results show that stylistic variation in model-generated data can be discovered in an unsupervised way, and made explicit, providing a source of both control and improved reasoning performance.

Ioana Marinescu, Eric Karl Oermann, Kyunghyun Cho

NYU · NYU & NYU Langone Health

Stylistic Fidelity Style Mathematical Reasoning Benchmarks Reasoning Skills Teacher Unsupervised

Sep 30, 2026cs.LG

Exploring More, Reasoning Better: Stepwise Risk-Sensitive GRPO for Diffusion Language Models

Diffusion large language models (dLLMs) generate text by denoising a sequence or successive blocks, allowing several tokens to be revealed in parallel. Reinforcement learning with verifiable rewards (RLVR) reuses terminal feedback across these decisions, even as their conditioning context changes. We propose stepwise risk-sensitive GRPO (StepRS-GRPO), which varies the risk coefficient of the group-advantage transformation across denoising states while retaining the underlying trainer. For binary rewards, we show that this transformation is exactly a prompt- and state-dependent rescaling of centered outcome advantages. A capability-based calibration suggests a coefficient scale, while endpoint and interpolation ablations guide schedule selection. Across multiple dLLM backbones and mathematical reasoning benchmarks, StepRS-GRPO improves both pass@1 accuracy and pass@k coverage over centered GRPO, while increasing answer diversity. In our ablation studies, mass-matched controls support the contributions of state allocation and schedule direction, and the gains persist after matching the root mean square (RMS) of the advantages to that of centered GRPO. Reasoning-trace diagnostics further show that the diversity gains from StepRS-GRPO extend beyond final-answer strings.

Yue YU, Bowen Zuo, David Crandall +2

Indiana University · University of California, Riverside

Diffusion Language Models Large Language Model Reinforcement Learning Reward Gradients Mathematical Reasoning Benchmarks Transform Concepts

Sep 29, 2026cs.CL

Correct Answers, Invalid Traces: What Verifiable Grade-School Math Reveals About Chain-of-Thought Traces

Chain-of-thought traces are widely read as records of how models reach their answers, informing debugging, agent auditing, and claims about reasoning. Testing this interpretation is difficult because natural-language thinking traces are rarely mechanically verifiable. We revisit it in iGSM, a synthetic grade-school mathematics benchmark designed to study thinking traces and used to support claims of learned reasoning and planning. Crucially, iGSM exposes the exact quantities and dependencies that a correct solution should use, allowing generated traces to be checked programmatically step by step and enabling us to test whether correct answers are reliably accompanied by valid traces. We first evaluate models trained exclusively on valid, minimal traces. Answer correctness and trace validity nearly coincide in distribution but decouple out of distribution: on the hardest instances, 31.6% of correct answers have invalid traces, over half of which pass all syntactic and arithmetic checks but fail semantic dependency checks. We then intervene on trace supervision. Non-minimal training traces induce non-minimal outputs, while re-asking the same problem with a different query reveals computations inherited from the original query, weakening minimality as evidence of selective planning. Shuffling tokens in 10% of training trace sentences preserves near-clean accuracy even out of distribution despite no trace passing verification. Swapped training traces likewise retain high in-distribution accuracy. We discuss the implications of these findings for chain-of-thought monitoring and interpretation in the context of AI safety.

Ratish Puduppully, Pranabendu Misra, Paarth Iyer +3

IT University of Copenhagen · Chennai Mathematical Institute · Indian Institute of Technology Jammu +1

Reasoning Traces Chain-of-Thought Reasoning Mathematical Reasoning Benchmarks Answer

Sep 29, 2026cs.CL

From Dissonance to Orchestration: Teacher Intervention in On-Policy Distillation

On-policy distillation (OPD) trains a student on its own reasoning trajectories using feedback from a stronger teacher. Teacher interventions can improve these trajectories, but also change the distribution on which the student learns. Our controlled studies show that rollout quality alone is an incomplete criterion for allocating teacher guidance. Deeper intervention yields diminishing gains in rollout accuracy while increasing off-policy load. In a training probe with a restricted rollout horizon, peak student accuracy and performance retention favor different intervention strengths. The preferred intervention depth and placement also vary across benchmarks. These findings motivate MAESTRO, which uses local policy disagreement to jointly adapt when the teacher takes over and how long it generates. Its {policy disagreement score} combines teacher-weighted candidate coverage with local distribution similarity and is aggregated within reasoning paragraphs. Across eight mathematical reasoning benchmarks, MAESTRO achieves the highest macro-average accuracy among the compared methods for both 0.6B and 1.7B Qwen3 students, with the 1.7B student leading on every benchmark. MAESTRO also reduces average training response length by 67.3% relative to standard OPD. The code is available at https://github.com/yhao-wang/MAESTRO.

Yuhao Wang, Ruiyang Ren, Yinan Zhang +3

Nanyang Technological University, Singapore · Baidu Inc.

Teacher Mathematical Reasoning Benchmarks Orchestration

Sep 27, 2026cs.LG

Selecting Diverse SFT Traces Improves Post-RL Generalization

Verified solutions are not equally useful for preparing reasoning models for reinforcement learning (RL). We present a comprehensive study of route diversity, the variation in the sequences of reasoning steps in supervised fine-tuning (SFT) data, and propose a lightweight, rule-based fingerprint to select for it. From one pool at one budget, with matched training recipes and checkpoints, selecting diverse rather than similar routes improves post-RL problem coverage across puzzles and mathematics, including on problems harder than those seen in either training stage. In synthetic experiments, route-diverse SFT improves OLMo3-7B's pass@8 by 16.9 points on environments held out from SFT. In a single-model condition, where one model writes every candidate, diverse selection gains up to 6.2 points of mean pass@8 across 10 mathematics benchmarks. Pre-RL diagnostics suggest why: diverse SFT can produce both successful and failed attempts on more prompts despite slightly lower mean accuracy, giving group-relative RL more prompts with a learning signal. On 3 open-source corpora, our CPU-only selector, without model calls, outperforms more expensive alternatives in every comparison of mean post-RL performance. These results identify reasoning-route diversity as a practical criterion for selecting SFT data that better prepares models for RL.

Dylan Zhang, Mingyuan Wu, Jinning Li

University of Illinois Urbana-Champaign, work done at Google · Google

Supervised Finetuning Reasoning Paths Mathematical Reasoning Benchmarks Group-Based Reinforcement Learning

Sep 27, 2026cs.AI

What Shared Prefixes Hide: Trajectory Dropout for On-Policy Distillation

On-policy distillation (OPD) trains a student model on its own trajectories using dense token-level feedback from a stronger teacher model. Since each update is conditioned on the reasoning prefix already generated by the student, the prefix also shapes how effectively teacher feedback is converted into learning. We find that shared prefixes can lead to weak token-level updates, a phenomenon we call Prefix-Induced Supervision Attenuation (PISA). This attenuation arises in two common cases. (i) High student confidence can weaken corrective gradients even when the teacher disagrees. (ii) Tokens that rely on earlier reasoning can receive learning signals as weak as those for simple local continuations. To solve this problem, we propose Trajectory Dropout, a simple training-time intervention that exposes these weakened signals. The student first performs a standard full-context rollout to generate a complete trajectory. During training, we randomly drop a certain proportion of the student's reasoning trajectory, while the teacher continues to observe the complete trajectory for token-level supervision. This intervention strengthens corrections for overconfident predictions and introduces additional supervision at prefix-sensitive positions. Trajectory Dropout consistently improves average performance across teacher--student model pairs of different scales and six mathematical reasoning benchmarks, while also yielding gains on two out-of-domain benchmarks. It can also be flexibly integrated into existing OPD variants with negligible computational overhead, further improving their performance. These results demonstrate that Trajectory Dropout provides a simple mechanism for strengthening token-level supervision across model scales and OPD objectives.

Zizhuo Lin, Quanling Liu, Yi Yang +1

Zhejiang University

Teacher Process-Level Supervision Mathematical Reasoning Benchmarks

Sep 24, 2026cs.CV

Multimodal Thinking with Renderable Programs

Current vision-language models (VLMs) excel at visual content understanding and text-based reasoning, yet their structure limits the advancement of incorporating images into the reasoning chain. Though Omnimodal models have made efforts in unifying text and image generation, they focus on visual tasks in the open-domain, lacking tractability due to rasterized or latent representations of images. We introduce SVGLM, a framework that uses scalable vector graphics (SVG) primitives to connect text and image in reasoning tasks. We exploit the duality of SVG as both image description and text instructions, yielding a more compact, interpretable solution to equip general VLMs with the capability of generating images within the reasoning process. We provide a large curated dataset of SVG-based image editing dataset, as well as the paradigm to tune open-source VLMs. Experiments on a mathematical reasoning benchmark demonstrate that SVGLM achieves strong SVG generation power as well as think-with-image intelligence. Our results highlight SVG as a suitable medium for building more robust digital domain agents, bridging the gap between text-based thinking and pixel-based images.

Sunli Chen, Ding Zhong, Ziqiao Ma +6

University of Massachusetts Amherst · University of Michigan · University of Illinois Urbana-Champaign +1

Scalable Vector Graphic Mathematical Reasoning Benchmarks

Sep 23, 2026cs.LG

When and Where to Trust the Teacher: Unifying On-Policy Distillation and GRPO through Entropy-Calibrated Credit Assignment

Reinforcement learning with verifiable rewards (RLVR) supervises mathematical reasoning through final-answer correctness, but provides little guidance on individual tokens. On-policy distillation (OPD) supplies dense feedback on student-generated responses, yet teacher preference need not reflect correctness. Recent hybrids combine OPD and verifier-derived advantages or reweight task credit using teacher ratios. However, teacher guidance enters after verifier-based group normalization, and token reweighting need not preserve the total task credit assigned to each response. We introduce Unified Entropy-Calibrated Credit Redistribution for GRPO (UECR-GRPO), which integrates verifier and teacher signals within a single GRPO-style update at both the response and token levels. \emph{Path-Utility Unification} (PUU) combines verifier reward and a teacher-to-anchor path log-ratio in a single KL-regularized objective. Its on-policy implementation uses a length-normalized teacher score and combines both rewards before group normalization and PPO clipping, allowing teacher evidence to influence the response ranking. \emph{Entropy-Calibrated Redistribution} (ECR) then uses the signed teacher--old-policy token gap to redistribute the verifier-derived component. Full-vocabulary teacher entropy attenuates uncertain guidance, while a response-wise zero-sum projection preserves the total task credit and its token-wise sign before clipping. Across five mathematical reasoning benchmarks, UECR-GRPO achieves average

\mathrm{Avg@12}

accuracies of 17.21% and 65.09% with Qwen3-1.7B and Qwen3-4B students, respectively, exceeding the strongest baseline at each scale by 0.89 and 0.56 percentage points.

Jie Zhang, Jingxiao Yang, Zhehao Huang +2

Shanghai Jiao Tong University · Zhejiang University

Reinforcement Learning With Verifiable Reward Teacher Verifiable Rewards Mathematical Reasoning Benchmarks

Sep 23, 2026cs.CL

Planned Test-Time Scaling with Coordinated Reasoning Paths

Test-time scaling with parallel branches is widely adopted to improve performance on challenging reasoning tasks. The predominant approach, repeated sampling, draws branches independently from a single policy, which can produce redundant attempts and thereby limit the gains from additional inference compute. To address this limitation, we propose Planned Test-Time Scaling (PTTS), which replaces independent sampling with a coordinated joint policy: a planner generates a solution outline for each branch, steering the branches toward distinct reasoning paths, and an executor produces a full solution conditioned on each outline. Formally, we show that PTTS strictly generalizes repeated sampling and, in a stylized setting, provably promotes coverage of complementary reasoning modes and yields better pass@k scaling. We instantiate PTTS on top of strong reasoning models, keeping them fixed as executors while replacing repeated sampling with PTTS inference to further enhance test-time scaling. Concretely, we develop two variants: PTTS-ZS prompts a model to jointly generate outlines for all branches in a single autoregressive pass, while PTTS-RL directly optimizes the planner against the pass@k reward using truncated execution rollouts for efficient training and a sharper reward signal. Across five mathematical reasoning benchmarks with Qwen3-1.7B and 4B, PTTS-ZS improves pass@64 over repeated sampling by up to 6.7 points, while PTTS-RL further increases the gain to up to 13.4 points. Further analysis indicates that broader coverage of distinct reasoning paths contributes to these gains. Overall, PTTS provides a general framework for improving test-time scaling by coordinating reasoning branches, with zero-shot and trainable instantiations that yield substantial performance gains.

Xueqing Wu, Langxing Bai, Hritik Bansal +5

University of California, Los Angeles · Amazon

Test-Time Scaling Test Time Reasoning Paths Mathematical Reasoning Benchmarks

Sep 22, 2026cs.CL

Giving Credit Where It's Due: Redundancy-Aware Learning for Efficient Reasoning

Large reasoning models can produce correct yet unnecessarily long reasoning traces. Existing methods improve reasoning efficiency with trajectory-level objectives or local token- and step-level signals, but rarely model inter-step semantic dependencies. This limits their ability to distinguish redundant steps from those that support later deductions, making it harder to shorten reasoning without sacrificing accuracy. We introduce RECAP (REdundancy-aware Credit Assignment via Propagation), which addresses this limitation by assigning credit where it is due based on both a step's downstream role in the reasoning structure and its contribution to solving the problem correctly. We define structural responsibility to capture the step's downstream role by measuring how strongly later reasoning depends on it, using credit propagated backward from the final-answer node through an outcome-independent, LLM-annotated semantic dependency graph. However, a step can have high structural responsibility yet steer the reasoning away from the correct solution. RECAP therefore introduces step efficacy to measure answer-directed progress through changes in gold-answer log-likelihood as each step is added. Together, these signals reshape rollout-level GRPO advantages into step-specific updates. RECAP requires neither a separately trained process reward model nor preconstructed concise trajectories. Across two 7B models and four mathematical reasoning benchmarks, RECAP improves the accuracy-efficiency trade-off. On Qwen2.5-Math-7B, it improves pass@1 by 2.0-3.7 percentage points while reducing reasoning tokens by 8%-31% relative to GRPO across all four benchmarks. Analysis suggests these savings reflect fewer reasoning operations and less dead-end reasoning, rather than more compact expression.

Yuqing Zhou, Hong Wang, Manqing Mao +8

George Mason University · Amazon, Inc

Reasoning Traces Mathematical Reasoning Benchmarks Step-Level Credit Semantic Relationships

Sep 22, 2026cs.AI

Math Reasoning in LLMs is Organized by Approach, Not Topic

Mathematical reasoning benchmarks are typically organized by topic, but language models may organize their internal computation by reusable reasoning approach instead. In this paper, we investigate whether open math-capable LLMs organize internally by topical sub-skill or by reasoning approach, and we present evidence that the approach is the key. We introduce a generation-replay protocol: a model first generates a solution, after which we replay the exact prompt-plus-generation trajectory and extract activation-importance signatures over the reasoning tokens. We cluster these signatures without supervision across eight models and five mathematical reasoning sources, then evaluate the recovered structure with structural, semantic, and intervention tests. Across all 40 model-source cells, the recovered clusters outperform matched-size random baselines. Two independent frontier-LLM judges find approach-level coherence in 77-82% of real clusters versus 6-11% in within-source controls, and topic-pure clusters usually receive labels finer than the topic itself. In approach-controlled prompting, changing the requested reasoning approach shifts cluster assignment in seven of eight model conditions, whereas paraphrases largely preserve it. These results indicate that math-capable LLMs organize internal mathematical computation by reasoning approach rather than benchmark topic. The implication is that topic-stratified benchmarks and topic-balanced training corpora can still miss the axis that matters: even deliberately topic-balanced corpora may remain imbalanced over reasoning approaches.

Sajad Goudarzi, Samaneh Zamanifard, Moloud Nasiri +1

Clemson University

Mathematical Reasoning Benchmarks Large Language Model Tool Use Text Corpora Virtual Cell

Sep 21, 2026cs.LG

iSDFT: Information-Proximal Self-Distillation for Continual Learning in LLMs

On-policy self-distillation fine-tuning (SDFT) learns new skills from demonstrations while reducing forgetting, but it always distils toward the full demonstration-conditioned teacher. This fixes teacher influence at the full-teacher endpoint, providing no control over how much demonstration information should be transferred at each prediction state. We introduce Information-Proximal SDFT (iSDFT), which instead treats the teacher as a budgeted source of information. At each token, iSDFT selects the distribution closest to the current student that satisfies a prescribed teacher-information constraint, yielding a closed-form exponential target with a locally determined tilt. To control cumulative drift, we further anchor the student to its frozen base policy. Across four heterogeneous LLM backbones and two specialisation tasks, iSDFT improves vanilla SDFT in 7 of 8 model-task settings and matches it in the remaining one. It also provides tighter retention on the original SDFT benchmark suite, with 73% of evaluations remaining within 0.5 points of the base model versus 52% for the strongest baseline, while achieving the largest mean improvement on all ten additional mathematics, coding, and competition-mathematics benchmarks. These results show that controlling how much and when teacher information is introduced improves specialisation while preserving broader capability.

Ahmed Khaled Khamis, Xiaotong Ji, Hassan Jaber +4

Imperial College London · Politecnico di Torino · UCL Centre for AI

Self-Distillation Framework Self-Teacher Teacher Continual Learning Mathematical Reasoning Benchmarks Interatomic Potentials

Sep 16, 2026cs.CL

Voice of Reason: Reinforcement Learning for Spoken Math

Speech language models enable richer spoken interactions between humans and machines than cascaded systems, allowing access to paralinguistic information and lower latency. However, their accuracy on mathematical reasoning benchmarks has lagged behind those of text models. Reinforcement learning (RL) with verifiable rewards has been instrumental in extending text models' capabilities for solving complex problems and limiting hallucinations. In this work, we explore applying RL to the GLM-4-Voice speech model (Zeng et al., 2024) to bridge the gap between textual and spoken mathematical problem solving. We first adapt the model to the domain using supervised fine-tuning on synthesized spoken question-answering data. We then show that, even without extra reasoning tokens, RL improves the accuracy on GSM8K beyond levels previously achieved for speech models only with supplementary reasoning traces. When combined with existing streaming reasoning techniques, we show further gains to 74.8% free-form accuracy. This establishes a new state-of-the-art for mathematical spoken abilities with speech-native models.

Timothée Weisselberger, Edouard Graves, Alexandre Défossez

Kyutai Paris, France · Kyutai, Gradium Paris, France

Speech Language Models Mathematical Reasoning Benchmarks LLM Reasoning Strategies Reasoning Skills Utterances Question-Answer Pairs

Sep 14, 2026cs.CV

ChitraMiti: Benchmarking Visual Grounding and Modality Reliance in Bengali Geometric Reasoning

Evaluation of vision-language models (VLMs) for multimodal mathematical reasoning remains limited for low-resource languages and for geometry problems that require reading a diagram and a question together. We introduce ChitraMiti-12.8k, a synthetic benchmark of 12,874 Bengali planar geometry problems paired with structured 15-attribute descriptions, and NCTB-500, a complementary set of 500 diagrams manually extracted from Bengali school textbooks. Using a three-phase protocol that separates diagram-only, diagram-plus-description, and description-only inputs, we show across five open-weight and closed-source VLMs that description-only performance is statistically indistinguishable from diagram-plus-description performance, establishing structured descriptions as a sufficient textual proxy for controlled evaluation. Despite this, models remain poor at cross-modal verification, frequently misled by a swapped spatial relation even when they answer the unmodified item correctly. We further evaluate supervised adaptation on ChitraMiti-12.8k, finding that fine-tuning improves performance on both ChitraMiti-1k and NCTB-500, although a substantial gap to the strongest zero-shot model remains. Together, ChitraMiti-12.8k, NCTB-500, and our evaluation protocol offer a standardized way to study Bengali multimodal geometry reasoning and, more broadly, whether VLMs actually check their text against what they see. Our dataset and code are publicly available on Hugging Face at https://huggingface.co/datasets/RaiyanKhaan/ChitraMiti.

Khan Raiyan Ibne Reza, Sanjana Aktar Maria, Sumaiya Tabassum Nimi +1

North South University

Multimodal Reasoning Benchmarks Mathematical Reasoning Benchmarks Bangla Diagrams Modalities

Sep 9, 2026cs.AI

Which Tokens Should SFT Actually Learn? A Token-Trimming Perspective on Mathematical Reasoning

Supervised fine-tuning (SFT) applies a uniform cross-entropy loss to all target tokens, even though different tokens provide unequal learning signals for mathematical reasoning. This uniform treatment can over-sharpen already mastered tokens while amplifying learning pressure on uncertain, low-confidence tokens, leading to suboptimal training dynamics. We propose Trimmed Logit-Gap SFT (TrimSFT), a simple token-level reweighting method that scales the SFT loss according to the logit gap between the gold token and its strongest competitor. TrimSFT trims supervision away from both extremes: tokens already mastered (large logit gap) and tokens weakly supported by the current model (small or negative logit gap), concentrating learning within an intermediate logit-gap region between them. We instantiate this principle with a Gaussian weight centered at margin m with bandwidth τ, requiring no reference model or additional forward pass. We evaluate TrimSFT on six base models from the Llama, Qwen, and DeepMath families across five mathematical reasoning benchmarks. TrimSFT consistently improves over standard SFT, achieving the best average performance on five out of six models, with gains of up to +26.9 points over SFT on MATH500. Further analyses show that the bandwidth τ matters more than the exact margin location, and that half-trim variants that remove supervision pressure from only one side yield inferior trade-offs. A token-level logit-gap distribution analysis suggests that TrimSFT reshapes model confidence in a more balanced way than uniform SFT or monotonic reweighting methods. These results suggest that reasoning SFT can benefit from trimming both extremes rather than treating all tokens uniformly.

Yaning Jia, Chunhui Zhang, Wenxuan Xu +3

Dartmouth College · Carnegie Mellon University

Supervised Finetuning Mathematical Reasoning Benchmarks Mathematical Reasoning Visual Token Pruning Cross-Entropy Losses

Sep 8, 2026cs.LG

Difficulty-Adaptive Tree-Structured Policy Optimization for Expanding Reasoning Coverage in RLVR

Reinforcement Learning with Verifiable Rewards (RLVR) has been central to the recent success of Large Reasoning Models. However, while RLVR significantly improves single-sample accuracy, it often fails to expand the model's intrinsic reasoning coverage (pass@k) due to limited exploration during training. To address this, we optimize the structural design of train-time rollouts to enhance pass@k. Our analysis identifies three key design principles: (1) difficulty-adaptive rollout can play an important role in expanding pass@k, beyond serving as an efficiency heuristic; (2) tree-based rollout outperforms parallel sampling in discovering correct answers; and (3) sentence-entropy-guided forking overcomes the localization phenomenon of token-level branching to maximize semantic diversity. Building on these insights, we propose DATPO (Difficulty-Adaptive Sentence-entropy-guided Tree-structured Policy Optimization). DATPO integrates difficulty-adaptive tree search with a sibling-diversity advantage term, explicitly promoting semantic diversity to expand reasoning coverage during training. Experiments on mathematical reasoning benchmarks demonstrate that DATPO outperforms baselines especially in pass@k, which directly translates to superior test-time scaling performance.

Youngjun Yu, Sanghwan Jang, Hwanjo Yu

Pohang University of Science and Technology (POSTECH)

Reinforcement Learning With Verifiable Reward Autoregressive Rollout Mathematical Reasoning Benchmarks

Sep 1, 2026cs.AI

Learning What to Practice: Diagnosis-Guided Self-Evolution for Language Models

Self-play supports the self-evolution of language models, but solver performance can plateau or decline across rounds without guidance. Existing unguided methods typically use difficulty, learnability, or diversity signals to keep questions challenging and varied, without identifying which unresolved reasoning weaknesses to target. Existing guided methods rely on external task resources such as human examples, document corpora, or specified difficulty targets. We introduce DiagEvo, which guides question generation using the solver's failure history from self-play, without external task resources. Its diagnostician extracts recurring error causes and stores them in an error-cause memory. The memory groups related causes under skill nodes and tracks each as Active or Mastered according to self-consistency on targeted questions. The challenger uses these states and recurrence counts to balance cause-targeted generation with free exploration. Double-confidence filtering retains intermediate-difficulty questions only when the most common solver answer has a clear vote lead. With the default 4B diagnostician, DiagEvo outperforms all baselines in mean accuracy across nine benchmarks for each solver: Qwen3-4B, Qwen3-8B, and OctoThinker-8B. On Qwen3-8B, DiagEvo reaches 72.3% mean accuracy across five mathematical reasoning benchmarks, 4.5 percentage points above R-Zero. Its overall mean accuracy across nine benchmarks is 57.4%, 3.5 percentage points above SPICE. Ablations show that mixed generation, memory-state updates with cross-state stitching, and double-confidence filtering contribute to these gains.

Xincheng Wei, Yifan Ding, Fucheng Xiong +6

The Chinese University of Hong Kong, Shenzhen · Meituan, LongCat Team · Peking University +2

Self-Evolution Mathematical Reasoning Benchmarks

Aug 13, 2026cs.AI

Numeracy in Large Language Models: Fundamental Limitations and Paths to Improvement

Large language models (LLMs) achieve strong results on mathematical reasoning benchmarks yet remain unreliable on elementary numerical tasks, including magnitude comparison, large-integer arithmetic, fractions, and scientific notation. This survey examines basic numerical understanding as a capability distinct from high-level mathematical reasoning. We propose the Numerical Grounding Framework (NGF), which decomposes numeracy into Representational Grounding (RG), mapping numeral forms to value, magnitude, and equivalent representations, and Procedural Grounding (PG), executing arithmetic operations in accordance with their mathematical definitions. Using NGF, we organize recent diagnostic benchmarks, failure modes, structural explanations, and mitigation strategies. We review evidence concerning tokenization, positional encoding, embedding geometry, and pretraining-data distribution. We also apply NGF in a coordinated evaluation of three frontier model families across Number Cookbook, NumericBench, and GSM-Symbolic, comparing atomic, contextual, and reasoning-assisted numeracy. Architectural interventions such as digit-aware tokenization and Abacus Embeddings can improve models trained from scratch but are generally unavailable to users of pretrained systems, for whom supervised fine-tuning, reasoning scaffolds, and external tools are more practical. We conclude with deployment recommendations and research directions for more reliable numerical behavior in foundation models.

Aoxin Ni

Department of Computer Science University of Chinese Academy of Sciences

Mathematical Reasoning Benchmarks Fundamental Limits Improvement Tokenizer Path

Aug 10, 2026cs.AI

Mismatch Matters: On-Policy Distillation Beyond Token Agreement

On-policy distillation (OPD) has emerged as a core component of modern LLM post-training pipelines, yet we reveal a failure mode: degenerate agreement, where students exploit repetitive loops to achieve near-perfect token agreement with the teacher despite globally flawed responses. We therefore shift our focus from agreement to teacher-student mismatch, and find that mismatch tokens can be mainly categorized into two types: student-excess tokens and student-deficit tokens. Student-excess tokens are generated by the student but assigned near-zero probability by the teacher; their log-ratio corrections grow unbounded and destabilize the update. Student-deficit tokens, in contrast, are preferred by the teacher but rarely sampled by the student; their absence blocks the transfer of the teacher's reasoning patterns. To tackle these mismatch directions, we propose TIDE (Token-level Independent Deficit-Excess correction), which applies bounded Hellinger shaping to suppress the most severe sampled excesses and an analytic teacher top-

K

injection to restore deficient probability mass without requiring deficit tokens to be sampled. Across mathematical reasoning benchmarks with multiple Qwen3 teacher-student pairs, TIDE consistently outperforms standard OPD and recent token-selection and reward-shaping baselines. Moreover, the gains of TIDE are more pronounced under strong teacher-student mismatch, where it improves Avg@8 from 6.9% to 20.3%, reduces average response length by a factor of 3.6, and substantially reduces formatting failures. Code is available at https://github.com/yzc-666/TIDE

Zichao Yu, Chengzhi Yu, Shengze Xu +4

The University of Hong Kong · University of Science and Technology of China · The Chinese University of Hong Kong

Teacher Mismatch Mathematical Reasoning Benchmarks

Aug 10, 2026cs.LG

Privileged Solutions or Context-Induced Teacher Behavior? Dissecting On-Policy Self-Distillation

On-Policy Self-Distillation (OPSD) is commonly interpreted as the transfer of privileged information: a teacher observes the verified solution to the target problem and supervises the student's trajectory. However, this interpretation conflates two effects. The reference solution not only reveals the answer to the current instance but also changes the context under which the teacher provides token-level supervision. We investigate the role of target-specific privilege with

\mathrm{OP}^{2}\mathrm{SD}

(On-Policy Self-Distillation from Other Problems), which replaces the paired reference with a problem and solution from a different example, while preserving the student rollout, teacher, and distillation objective. Across three models and three mathematics benchmarks,

\mathrm{OP}^{2}\mathrm{SD}

improves over the base model, remains competitive with OPSD. The success of

\mathrm{OP}^{2}\mathrm{SD}

implies that OPSD gains do not necessarily come from access to the reference solution, and that the teacher's context-induced behavior is an important factor.

Yuki Ichihara, Naoto Iwase, Mohammad Atif Quamar +1

Mohamed bin Zayed University of Artificial Intelligence · Nagoya University · RIKEN AIP

Teacher Privileged Context Mathematical Reasoning Benchmarks Agent Behavior Modeling

Aug 9, 2026cs.LG

PAST: Privileged Adaptation from Complete Student Trajectories for On-Policy Self-Distillation

On-policy self-distillation (OPSD) uses a privileged teacher to supervise a reasoning model on prefixes sampled from its own rollouts. Yet each rollout also reveals how the student's response unfolds and whether it succeeds, student-specific hindsight that standard OPSD does not use to form the teacher. We introduce Privileged Adaptation from Student Trajectories (PAST), which treats each completed student trajectory as additional privileged information for the OPSD teacher while leaving the student's distillation prefixes unchanged. PAST preserves the student's next-token distribution on correct trajectories and uses failed trajectories to adapt the teacher toward verified success under student-proximity regularization. We characterize what such a trajectory-conditioned teacher can transfer to a prefix-only student. Forward-KL distillation projects the teacher distributions to their conditional arithmetic mean given the prefix. This projection separates trajectory-specific variation that remains privileged from the mean policy shift available to the student. For correct trajectories, the unclipped population objective also has the frozen student as an ideal distributional fixed point. Across three mathematical reasoning benchmarks, PAST improves the Avg@12 macro average over Vanilla OPSD by 5.6 percentage points. A

2\times2

factorial study shows gains from both complete-trajectory access and teacher adaptation, while trajectory removal and shuffling confirm that the adapted teacher uses the matching hindsight context.

Yangyang Feng, Zhuoyan Feng, Junlan Chen

1The Hong Kong University of Science and Technology (Guangzhou) · 2Sun Yat-sen University

Single Trajectory Mathematical Reasoning Benchmarks Text-Only Adaptation

Aug 6, 2026cs.CL

RP-OPSD: Reasoning-Pivot-Guided On-Policy Self-Distillation for Multilingual Reasoning Transfer

Multilingual reasoning transfer is crucial for extending reasoning capabilities of large language models (LLMs) beyond high-resource languages. On-policy self-distillation (OPSD) and its variants have emerged as a promising paradigm, providing dense token-level supervision on student-generated rollouts, yet their objectives do not explicitly prioritize reasoning signals most critical to cross-lingual transfer. We characterize that target-language reasoning comprises the generation of both surface text and reasoning pivots, which are decisions that advance or redirect the reasoning process and shape subsequent inference. This motivates concentrating privileged distillation around such pivots. We therefore propose RP-OPSD, Reasoning-Pivot-guided On-Policy Self-Distillation, using the distributional shift between matched teacher views with and without an English reference solution as an operational proxy to guide privileged distillation and reference anchoring. Experiments on mathematical reasoning benchmarks covering 17 languages and multiple difficulty levels show that our method outperforms strong multilingual reasoning baselines and OPSD variants. Further analysis reveals that RP-OPSD concentrates privileged distillation on reasoning-control and problem-condistioned state-update tokens, while downweighting it for tokens that mainly support surface realization. Our code is available at https://github.com/NJUNLP/RP-OPSD.

Xinye Wang, Junxiao Liu, Shujian Huang

National Key Laboratory for Novel Software Technology, Nanjing University

LLM Reasoning Strategies Unsupervised On-Policy Self-Distillation Mathematical Reasoning Benchmarks Rp-Opsd On-Policy

Aug 3, 2026cs.AI

Is More Privileged Information Better? From Solution Traces to Problem-Solving Structure in Self-Distilled Reasoning

On-policy self-distillation (OPSD) improves reasoning by using a privileged view of a model conditioned on reference solutions to supervise a student view that observes only the question. However, the teacher-provided token-level targets may depend on reference-specific information unavailable at inference time. We propose Problem-Space-Guided OPSD (PS-OPSD), which replaces the complete solution with trajectory-grounded guidance describing the initial state, goal conditions, constraints, and a selected state-transition path. The student rollout and OPSD objective remain unchanged. Across three mathematical reasoning benchmarks and model scales ranging from 1.7B to 8B, PS-OPSD achieves the highest aggregate question-only accuracy among the compared methods. Controlled experiments further indicate that guidance relevance and path coherence contribute to these gains, highlighting the representation of privileged information as an important design choice in OPSD.

Xuyang Zhao, Liting Zhang, Zichen Xu +4

1TMCC, College of Computer Science, Nankai University, Tianjin, China · 2Huawei Technologies Ltd., Beijing, China

Mathematical Reasoning Benchmarks Privileged Context Problem Solving Capability Guidance Traces

Aug 1, 2026cs.AI

TrAC: Trace-Conditioned Answer Consistency for Efficient Uncertainty Quantification in LLMs

Large language models (LLMs) can generate fluent reasoning traces that nevertheless lead to incorrect answers, making response-level uncertainty estimation important for abstention, human review, and adaptive compute allocation. Existing approaches generally fall into three categories: passive single-trace methods use token-level confidence signals, sampling-based methods compare multiple complete traces at higher generation cost, and active prefix-based methods probe partial traces to study answer stabilization or preference transitions. However, none actively re-elicits an answer from a completed reasoning trace to measure its consistency with and support for the original answer. To address this gap, we introduce Trace-Conditioned Answer Consistency (TrAC), a correctness-supervised uncertainty quantification framework that combines active and passive signals anchored to one completed reasoning trace. Its active component, Prefix-Conditioned Elicitation (PCE), re-elicits a short answer conditioned on the completed trace and represents both its consistency with the original answer and its token-level probabilistic support. Its passive component, Trace Uncertainty Profile (TUP), summarizes how token-level uncertainty evolves throughout the original generation without additional decoding. A lightweight head then integrates the two representations into a response-correctness score. Across five mathematical reasoning benchmarks and three LLM families, TrAC improves macro AUROC by 1.8% and reduces AURC by 3.4% relative to eight-sample self-consistency, while using one complete reasoning trace and a short cached answer probe. When eight samples are already available, augmenting sample consensus with re-elicitation further improves macro AUROC by 4.3% and reduces AURC by 8.3%, without additional full-trace generation.

Dahai Yu, Lin Jiang, Rongchao Xu +1

Florida State University Tallahassee, Florida, USA

Large Language Model Uncertainty LLM Reasoning Strategies Mathematical Reasoning Benchmarks Uncertainty Quantification Token-Level Uncertainty

Jul 30, 2026cs.CL

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B

Methods that make a language model plan, criticise and rewrite its own answer, reflect on mistakes, pick the best of several attempts, or debate with copies of itself nearly all make it generate far more text than a single chain of thought. Because generating more text raises accuracy by itself, a gain over one chain of thought does not show the method's idea is what helped. Wang et al. (2024) reported that a simple baseline, sampling the same question repeatedly and keeping the most common answer, often wins once budgets are comparable, but gave point estimates with no confidence intervals or significance tests. We rerun that comparison as a designed experiment: seven methods, open models of 1.5B, 3B and 7B parameters, two mathematics benchmarks, 150 questions each. We count every generated token, including those spent on critiques, reflections, debate turns and checking, and compare each method against repeated sampling at its own measured cost. All 36 comparisons are paired by question, with bootstrap intervals and multiplicity correction. No method is reliably better than repeated sampling at equal cost anywhere. Ten are reliably worse, all of them methods where the model inspects its own output, and all 18 self-inspection comparisons are negative. The two kinds of self-inspection part company as models grow. Choosing stops hurting: taking Best-of-N's eight samples and just counting the most common answer beats letting the model pick by 8.0 and 11.3 points at 1.5B, but only 2.0 and 1.3 at 7B, no longer distinguishable from zero. Rewriting does not recover: Self-Refine and a forced Reflexion stay 3.6 to 10.1 points below baseline at 7B. Reflexion as published never triggered its own retry on the smallest model. It judged itself correct every time and silently became a single chain of thought. We release code, prompts, all generations, and our verification scripts.

Iliya Mirzaei

Department of Computer Science Stony Brook University

Self-Referential Bias Reflections Refinement Mathematical Reasoning Benchmarks

Jul 30, 2026cs.AI

SVR: Self-Verifying Refinement via Joint Verdict-Confidence Reinforcement Learning for Adaptive Test-Time Compute

Scaling test-time computation can improve language-model reasoning, but uniform budgets waste computation on easy inputs, while verifier-guided refinement relies on external feedback. We introduce Self-Verifying Refinement (SVR), an oracle-free multi-turn reinforcement learning framework that learns to use self-verification as a compute-control policy. At each turn, the model produces a solution together with a discrete correctness verdict and a confidence score; it retains the current answer only when the verdict is Correct and confidence exceeds a threshold, and otherwise continues refinement using its own self-verification. Ground-truth correctness is used only to construct training rewards and is never exposed to the policy through refinement prompts or required at inference. SVR is trained with GRPO on fixed-horizon trajectories using rewards that promote solution correctness, calibration-aware self-verification, and stop-ready correct states; adaptive stopping is activated only at inference. On seven mathematical reasoning benchmarks with Qwen3.5-2B, SVR achieves a macro-average accuracy of 0.563 with only 2.99 inference turns on average. In the evaluated complete-system comparison, it exceeds standard GRPO, strong multi-turn baselines, and a fixed-budget oracle-guided score-feedback reference while requiring substantially fewer turns than fixed ten-turn inference. These results demonstrate that learned self-verification can serve as an effective internal control signal for answer retention and adaptive test-time compute allocation.

Hongyu Chen, Liang Lin, Guangrun Wang

1Sun Yat-sen University · 2Guangdong Key Laboratory of Big Data Analysis and Processing · 3X-Era AI Lab

Reinforcement Learning With Verifiable Reward Test Time Mathematical Reasoning Benchmarks

Jul 30, 2026cs.CL

LEEPS: Latent-Guided Explore-Exploit Prompt Sampling for Efficient RLVR in Large Language Models

Reinforcement learning with verifiable rewards (RLVR) improves the reasoning capabilities of large language models, but prompt groups with identical rollout rewards consume generation budget without effective learning signals. Pre-rollout prompt selection can reduce this waste by screening prompts before rollout generation. However, existing pre-rollout methods struggle to balance exploitation and exploration: repeatedly exploiting historically informative prompts can narrow training coverage, whereas broader exploration can lower the fraction of informative prompts. To address these limitations, we introduce LEEPS, a Latent-Guided Explore--Exploit Prompt Sampler that adaptively balances the reuse of previously observed informative prompts with continued exploration of uncertain ones. LEEPS partitions candidates into exploit and explore portfolios and adaptively allocates rollout budget according to their recent non-trivial ratios. It further uses representation-space neighbors and historical rollout outcomes to prioritize uncertain prompts likely to yield non-zero reward variance, thereby making exploration more targeted without additional rollouts. Across six mathematical reasoning benchmarks, LEEPS achieves the highest average score at both model scales, with relative gains of 2.6% and 3.7% over the strongest baseline for Qwen2.5-Math-1.5B and 7B, respectively, and generally improves faster during the training process. It also achieves the highest average score across the three evaluated OOD general-reasoning benchmarks at both model scales and adds only about 2 seconds of online sampling overhead per training step. Code is available at https://github.com/ShuangLiangX/LEEPS.

Shuang Liang, Haoyang Zhou, Yifan Gong +2

Renmin University of China · AntGroup

Reinforcement Learning With Verifiable Reward LLM Reasoning Strategies Mathematical Reasoning Benchmarks Prompt Engineering Sleep

Jul 30, 2026cs.AI

ReDiPPO: Reference-Guided Value Calibration and Discrepancy-Aware Token Reweighting for Mathematical Reasoning

Reinforcement learning has emerged as an effective paradigm for enhancing the mathematical reasoning capabilities of large language models. Among existing policy optimization methods, Proximal Policy Optimization (PPO) remains particularly appealing because its learned critic can, in principle, provide token-level credit assignment. However, in mathematical reasoning tasks characterized by long reasoning horizons and sparse outcome rewards, reliable token-level credit assignment remains challenging. The standard critic often fails to accurately evaluate intermediate reasoning states, resulting in noisy advantage estimates and suboptimal policy updates. In this paper, we propose ReDiPPO, a Reference-guided and Discrepancy-aware PPO framework for mathematical reasoning. ReDiPPO introduces a reference-guided critic that uses reference answers as training-time privileged signals to provide more accurate value estimation. Meanwhile, it retains a standard critic and quantifies the token-level reference-standard discrepancy between the standard value estimate and the reference-guided value estimate. This discrepancy serves as an indicator of difficult reasoning states and is used to reweight the corresponding token-level advantages during PPO optimization. Extensive experiments on diverse mathematical reasoning benchmarks demonstrate that ReDiPPO improves value-estimation accuracy and consistently outperforms strong policy optimization baselines, including PPO, DAPO, and GSPO, in final reasoning performance. Our code is available on https://github.com/cii030/ReDiPPO.

Zhenrong Zhang, Fei Wu, Jun Du +2

1IFLYTEK Research · University of Science and Technology of China

Mathematical Reasoning Benchmarks Mathematical Reasoning Credit Assignment Reweighting Discrepancy

Jul 30, 2026cs.LG

Kalman Meets Curriculum: Efficient Dynamic Prompt Selection for Adaptive RL Finetuning

Reinforcement learning (RL) finetuning significantly enhances the reasoning capabilities of large language models (LLMs), yet its effectiveness critically depends on selecting prompts of appropriate difficulty for the current policy. This is challenging because prompt difficulty evolves throughout training. Existing online methods therefore face a trade-off: evaluation-based approaches are accurate but expensive, while prediction-based approaches are efficient but typically assume stationary difficulty, making them ill-suited to RL's non-stationary training dynamics. To address these issues, we propose a Kalman-Guided Prompt Selection method (KGPS), which reformulates prompt selection as a dynamic state estimation problem rather than static difficulty prediction. KGPS models each prompt's latent success rate in logit space using a linear-Gaussian state-space model, with process noise coupled to the magnitude of policy updates so that uncertainty increases when the policy changes more substantially. A Kalman filter then maintains a calibrated Gaussian posterior over prompt difficulty, and prompts are selected by maximizing a posterior-expected training utility that favors intermediate-difficulty prompts while naturally revisiting uncertain ones. The resulting procedure is adaptive to policy drift and requires no additional rollouts beyond standard policy training. Extensive experiments across mathematics, planning, and geometry reasoning benchmarks, as well as multiple RL algorithms, show that KGPS consistently improves both final accuracy and rollout efficiency over strong baselines, establishing state-of-the-art performance among online prompt selection methods. For example, on DeepSeek-R1-Distill-7B, KGPS uses 83% fewer rollouts than DS while even improving the average performance by 0.12 point across six math reasoning benchmarks.

Haodong Zhu, Yangyang Ren, Yanjing Li +4

1Beihang University · 2Zhongguancun Academy · 3Nanyang Technological university +1

Automatic Prompt Optimization Mathematical Reasoning Benchmarks Curriculum Extended Kalman Filter

Jul 29, 2026cs.LG

ReCo: Reweighting GRPO Against Distributional Concentration

Group Relative Policy Optimization (GRPO) has become a standard reinforcement learning method for post-training language models. Recent work shows that GRPO can reduce the base model's reasoning capacity and underperform it in Pass@k when k is large, indicating reduced coverage of reasoning paths. We find that this reduction is associated with GRPO concentrating on responses that the base model already generates with high probability. We trace this concentration to two mechanisms in the GRPO update. At the response level, high-probability responses dominate the group gradient through repeated occurrence. At the token level, GRPO's importance ratio scales gradients, further reinforcing tokens that become more likely under the current policy. We propose ReCo, a reweighting method that addresses both effects. Response contributions are normalized by their expected occurrence within the rollout group, and the token-level importance ratio is replaced with a variance-based ratio that gives larger update scale to non-saturated decision points where alternative token choices remain plausible. Across Qwen2.5-Math-1.5B/7B and Llama-3.1-8B-Instruct on five mathematical reasoning benchmarks, ReCo improves Pass@k for large values of k and is comparable to GRPO for small values of k.

Junoh Park, Junseo Hwang, Wonguk Cho +1

Graduate School of Data Science, Seoul National University

Reweighting Mathematical Reasoning Benchmarks Token-Level Uncertainty Concentration

Jul 28, 2026cs.CL

Pass the Baton: Trajectory-Relayed On-Policy Distillation

On-policy distillation (OPD) grounds token-level supervision in the student's own trajectory, yet suffers from prefix failure: once the student commits to a wrong reasoning direction, all subsequent generation builds on this deviation, producing misdirected continuations that elicit unreliable supervision and waste compute. We identify a teacher-student continuation asymmetry on failed prefixes, where the teacher tends to redirect while the student continues along the original direction, and convert it into a label-free handoff trigger in Relay On-Policy Distillation (Relay-OPD). During training, Relay-OPD constructs relay trajectories by letting the teacher briefly take over at detected trigger points to produce a teacher leg, after which the student resumes and is optimized on the resulting trajectory. A limited relay budget concentrates intervention on critical early positions while limiting departure from the student policy. With a Qwen3-4B-Instruct-2507 teacher and Qwen3-0.6B/1.7B-Non-Thinking students on eight mathematical reasoning benchmarks, Relay-OPD achieves the best or second-best results on every benchmark, outperforming standard OPD by +5.73% and the strongest baseline FastOPD by +1.49% on average for 1.7B, with consistent gains at 0.6B. Training trajectory length is reduced by over 50%.

Haolei Xu, Xiaowen Xu, Haiwen Hong +5

1Zhejiang University · 2Yuvion Team, Alibaba Group

Teacher Mathematical Reasoning Benchmarks

Jul 13, 2026cs.CL

AdvancedMathBench: A Benchmark Suite for Advanced Mathematical Proof Generation and Verification

Large language models (LLMs) have achieved remarkable performance on high-school and competition-level mathematics, yet their capabilities on advanced mathematics remain poorly understood. Existing benchmarks, however, fall short in both scope and evaluation granularity: they provide limited disciplinary coverage and often rely on final-answer correctness or coarse judgments, leaving the validity of the reasoning process inadequately assessed. To bridge this gap, we introduce AdvancedMathBench, a benchmark suite designed to evaluate the reasoning capabilities of LLMs on advanced mathematical proofs. Its core generation benchmark, ProverBench, contains 245 problems spanning undergraduate (UG) and doctoral qualifying-exam (QE) levels. To reliably evaluate these proofs, we develop a dedicated automatic verification pipeline that is trained on large-scale expert annotations, produces both correctness verdicts and fine-grained analyses, and exhibits strong agreement with human experts on held-out proof trajectories. We further introduce VerifierBench, consisting of 888 model-generated proof trajectories paired with expert ground truth, to evaluate whether models can correctly judge proof validity and provide sound verification rationales. Experiments show that AdvancedMathBench remains challenging for frontier models. On proof generation, the best-performing model, GPT-5.5-xhigh, achieves only 64.5 and 48.9 on the UG and QE splits, respectively. On proof verification, the best model only attains a Balanced F1 of 65.1. Further analysis reveals a notable mismatch between proof generation and verification capabilities across models.

Lingkai Kong, Zijian Wu, Yuzhe Gu +11

Shanghai AI Laboratory · Shanghai Jiao Tong University · MMLab, The Chinese University of Hong Kong +1

Research-Level Mathematics Mathematical Reasoning Benchmarks Mathematical Reasoning Theorem Proving Large Language Model Benchmarks Data Generation

Jul 8, 2026cs.AI

MIRA-Math: A Benchmark for Minimal Information Requesting and Mathematical Reasoning

Mathematical reasoning benchmarks typically provide all facts needed to solve each problem, while interactive benchmarks often mix reasoning with tools, retrieval, and long-horizon dialogue. We introduce MIRA-Math, a benchmark for a narrower diagnostic capability: solving mathematical problems whose full latent state has a unique answer, but whose solver-facing view is missing exactly one necessary atomic fact. The solver must request the missing information in natural language under a strict budget and then integrate the returned fact into an exact final answer. A fixed constrained LLM responder sees only the dataset-provided atomic fact and must either offer the quoted fact when the request matches it, or decline otherwise. Thus, instance generation, typed hint specifications, validation, and final-answer verification are deterministic, while request metrics are measured under a fixed LLM-mediated responder channel. MIRA-Math contains 2{,}310 generated instances from 22 typed mathematical families spanning algebra, probability, linear systems, discrete structures, signal processing, Markov chains, circuits, interpolation, and numerical boundary-value problems. Experiments across frontier and small models show that request success and final-answer accuracy are separable: models may ask for the right fact yet fail the downstream computation, or fail before obtaining the canonical hint. We release generators, verifiers, prompts, run metadata, and dataset documentation to support reproducible evaluation of minimal information requesting in mathematical reasoning.

Charbel Al Bateh, Samer Saab

Department of Electrical and Computer Engineering Lebanese American University Byblos, Lebanon

Mathematical Reasoning Mathematical Reasoning Benchmarks

Jul 6, 2026cs.AI

LLM-as-a-Verifier: A General-Purpose Verification Framework

Scaling pre-training, post-training, and test-time compute have become the central paradigms for improving the capabilities of LLMs. In this work, we identify verification, the ability to determine the correctness of a solution, as a new scaling axis. To unlock this and demonstrate its effectiveness, we introduce LLM-as-a-Verifier, a general-purpose verification framework that provides fine-grained feedback for agentic tasks without requiring additional training. Unlike standard LM judges that prompt LLMs to produce discrete scores for candidate solutions, LLM-as-a-Verifier computes the expectation over the distribution of scoring token logits to generate continuous scores. This probabilistic formulation enables verification to scale along multiple dimensions: (1) score granularity, (2) repeated evaluation, and (3) criteria decomposition. In particular, we show that scaling the scoring granularity leads to better separation between positive and negative solutions, resulting in more calibrated comparisons. Moreover, scaling repeated evaluation and criteria decomposition consistently lead to additional gains in verification accuracy through variance and complexity reduction. We further introduce a cost-efficient ranking algorithm for selecting the best solution among candidates using the verifier's continuous scores. LLM-as-a-Verifier achieves state-of-the-art performance on Terminal-Bench V2 (86.5%), SWE-Bench Verified (78.2%), RoboRewardBench (87.4%), and MedAgentBench (73.3%). Beyond verification, the fine-grained signals from LLM-as-a-Verifier can also serve as a proxy for estimating task progress. We build an extension for Claude Code, enabling developers to monitor and improve their own agentic systems. Finally, we show that LLM-as-a-Verifier can provide dense feedback for RL, improving the sample efficiency of SAC and GRPO on robotics and mathematical reasoning benchmarks.

Jacky Kwok, Shulu Li, Pranav Atreya +6

Stanford University · UC Berkeley · NVIDIA Research

Large Language Models(Llms Feedback Mathematical Reasoning Benchmarks

Jul 1, 2026cs.CL

Know When to Stop: Segment-Level Credit Assignment for Reducing Overthinking

Reasoning language models frequently overthink: generating extended chains of behaviors such as hedging, approach abandonment, and self contradiction that consume tokens without improving answers. We show that these behaviors are not merely a consequence of length; even when controlling for response length, incorrect traces exhibit higher rates of unproductive self-reflection than correct ones. Addressing this requires identifying where self-reflection helps vs hurts, but obtaining these step-level annotations is costly. We observe that intermediate answer commitments within reasoning traces can provide a cheap proxy: by comparing each final answer candidate in the trace to the ground truth, we can determine whether subsequent reflection is productive without any additional supervision. Building on this insight, we propose DASH (Drift Aware advantage SHaping), which assigns segment-level credit based on whether each reasoning segment leads toward or away from correctness. On competition-level math benchmarks, DASH achieves the highest accuracy where overthinking is prevalent (AIME25: 50.8% vs. 45.4% GRPO) while reducing overthinking behaviors and achieving more productive self-correction than baselines.

Chia-Hsuan Lee, Sihui Dai, Mingyang Zhou +4

Capital One

Reasoning Traces Iterative Self-Correction Mathematical Reasoning Benchmarks Credit Assignment Step-Level Credit Commitment

Jun 29, 2026cs.AI

Does Verbose Chain-of-Thought Really Help? In-Distribution Evidence that Content, Not Length, Matters

Chain-of-thought (CoT) prompting improves LLM reasoning, but the source is contested: do the intermediate steps help because they carry useful semantic content, or because conditioning on more tokens buys extra computation before the model commits to an answer? We bring two lines of evidence to bear. First, in distribution: we repeatedly sample each model on the same question and pair a shorter with a longer of its own natural generations that follow the same reasoning plan, so nothing is rewritten and both traces are genuinely in-distribution. Across 25 models the extra tokens leave accuracy essentially unchanged for every independently-trained reasoner, and a blind analysis of the surplus tokens shows that what gain exists elsewhere tracks validation- and checking-content, not verbosity per se. Second, as a controlled intervention, we ask whether two traces expressing the same semantic content (the same facts, operations, and intermediate values, verified through directed acyclic graph equivalence) produce different outcomes when one is more verbose, using a dual-validator design across four targets and eight benchmarks with number-redacted completion and stratified bootstrap confidence intervals. Verbose traces do improve accuracy (25 of 32 benchmark-target cells are positive under at least one validator), but the effects are modest (typically 1-4 points) and depend on the quality of the verbose prose, not merely its length. Under maximum numerical redaction the effect is amplified (median 3.24x across four arithmetic benchmarks), and length-matched non-reasoning filler recovers none of it. Both lines converge: what matters is what the extra tokens do (the reasoning and validation content they carry), not how many there are, a picture neither a pure forward-pass-compute nor a pure semantic-content account fully explains.

Wenlong Wang, Fergal Reid

Fin AI Research

Chain-of-Thought Reasoning LLM Reasoning Strategies Mathematical Reasoning Benchmarks In-Distribution

Jun 29, 2026cs.IR

SABER-Math: Automated Benchmark for Information Retrieval Evaluation in Mathematics

As agentic AI systems tackle more complex mathematical tasks, they increasingly rely on information retrieval (IR) to search problem databases, theorem libraries, and educational resources. However, choosing the right retriever remains difficult, as it is infeasible to directly isolate its effect on downstream performance. On the other hand, existing retrieval-specific benchmarks often fail to capture fine-grained mathematical relevance, penalizing relevant documents. We address this gap by introducing SABER-Math, the first fully automated benchmark for evaluating mathematical IR without expert annotation. Starting from 283K high-school-level math problems with solutions, SABER-Math builds challenging reranking tasks in three steps: (i) first, LLMs extract concise solution summaries and mathematical topics for each problem; (ii) then, per-query relevant documents are discovered using ontology topic-based and lexical solutions-summary-based similarities, and (iii) finally, a Swiss-style LLM preference tournament produces fine-grained relevance ratings for the documents. We evaluate lexical retrievers, specialized mathematical retrieval systems, and recent embedding models. We find that while modern embedding models substantially outperform classical and math-specific baselines, even the strongest systems struggle in symbol-heavy domains like Algebra and Calculus. Importantly, we show that general-purpose IR benchmarks such as MTEB do not reliably predict mathematical performance, especially for recent embedding models, highlighting the need for math-specific retrieval benchmarks.

Nikolay Georgiev, Maria Drencheva, Kseniia Ibragimova +3

INSAIT, Sofia University “St. Kliment Ohridski” · ETH Zurich

Mathematical Reasoning Benchmarks Mathematics

Jun 22, 2026cs.LG

ReNIO: Reweighting Negative Trajectory Importance for LLM On-Policy Distillation

On-policy distillation (OPD) improves LLM reasoning by training a student model on its own generated outputs, but standard OPD treats all student-generated outputs (SGOs) equally regardless of their informativeness. We observe a consistent asymmetry in controlled filtering experiments: in both OPD and on-policy self distillation (OPSD), training only on incorrect SGOs outperforms training only on correct ones. Our further analysis suggests that models trained on correct-only SGOs tend to generate shorter reasoning traces and show weaker reflection behavior, while incorrect SGOs better preserve exploratory reasoning near the model's capability boundary. To exploit this signal without requiring full answer-containing rollouts, we introduce ReNIO, which Reweights Negative trajectory Importance for LLM On-policy distillation. By using the student-to-teacher probability ratio, ReNIO identifies pivotal tokens leading to wrong reasoning traces and aggregates their information into a normalized sample weight, inherently assigning larger weights to likely negative trajectories without observing the correctness of final-answer. Since Re-NIO only uses prefix-conditioned token probabilities, it preserves OPD's prefix training advantage over full-rollout reinforcement learning. Across both mathematical reasoning and code generation tasks, ReNIO improves both OPD and OPSD, with representative relative gains of up to 8.90% for Qwen3-1.7B and 10.00% for R1-Distill-Qwen-7B on mathematical reasoning benchmarks. Code repo: https://github.com/BDML-lab/ReNIO.

Chen Lin, Kedi Chen, Wei Zhang

East China Normal University · Shanghai Innovation Institute

Mathematical Reasoning Benchmarks Negative Results

Jun 10, 2026cs.AI

Architecture-Aware Reinforcement Learning Makes Sliding-Window Attention Competitive in Math Reasoning

The rapid progress of reasoning and agentic large language models (LLMs) has increased the demand for long-context inference, but self-attention (SA) scales quadratically with context length. To address this, we study SWARR (Sliding-Window Attention with Reinforced Adaptation for Math Reasoning), a practical recipe for adapting SWA models to mathematical reasoning. SWARR has two stages: (1) efficient conversion from a pretrained SA model to SWA with supervised fine-tuning (SFT), which avoids pretraining a new base model, and (2) policy adaptation with reinforcement learning (RL). We find that SWA still underperforms SA after SFT, and we hypothesize that this gap is caused in part by a data-architecture mismatch: most SFT data are prepared for SA models and may contain long-range dependencies that are difficult for SWA to model. Because on-policy RL optimizes self-generated trajectories under the SWA constraint, it can adapt trajectories to better match SWA. Experiments on mathematical reasoning benchmarks show that this recipe substantially narrows the gap between SWA and SA, recovering much of the accuracy lost during SWA conversion while preserving the efficiency benefits of linear-complexity attention. Our central contribution is the empirical finding that RL changes the conclusion one would draw from conversion and SFT alone about SWA's viability for math reasoning.

Kai Liu, Peijie Dong, Xinchen Xie +5

Shanghai Research Institute for Intelligent Autonomous Systems, Tongji University · Hong Kong University of Science and Technology (Guangzhou) · Shanghai AI Laboratory

Mathematical Reasoning Benchmarks Large Language Model Reinforcement Learning Kimi Delta Attention Rapid Adaptation

Jun 9, 2026cs.LG

N-GRPO: Embedding-Level Neighbor Mixing for Enhanced Policy Optimization

The success of Large Language Models in mathematical reasoning relies heavily on the generation of diverse and valid solution paths during the rollout phase. However, current rollout techniques face a fundamental trade-off: token-level sampling often yields redundant trajectories that differ only in rephrasing, while embedding-level methods utilizing random noise frequently disrupt semantic consistency. To resolve this, we introduce N-GRPO, a novel exploration strategy integrated into the Group Relative Policy Optimization (GRPO) framework. Rather than relying on token-level sampling or native embedding-level noise, our approach leverages Semantic Neighbor Mixing. This mechanism dynamically constructs input representations by mixing the embeddings of an anchor token and its nearest semantic neighbors, thereby injecting diversity while strictly adhering to the local semantic manifold. Experimental evaluations on the DeepSeek-R1-Distill-Qwen models across different sizes show that N-GRPO not only achieves consistent improvements over strong baselines on math reasoning benchmarks but also exhibits robust generalization capabilities on out-of-distribution tasks.

Xukun Zhu, Hang Yu, Peng Di +1

Zhejiang University · Ant Group

Mathematical Reasoning Benchmarks Flow-Grpo

Jun 9, 2026cs.CL

KCSAT-ML: Probing Reasoning Models with Nationwide-Cohort Human Difficulty

Math reasoning benchmarks have proliferated, yet most lack a per-item difficulty signal grounded in actual human performance. We introduce KCSAT-ML, a decade (2014-2025) of Korean College Scholastic Ability Test (KCSAT; Suneung) mathematics: 664 problems with a 339-item core set carrying official per-item error rates from nationwide cohorts of hundreds of thousands of examinees. We pair the benchmark with Difficulty-aligned Reasoning Gain (DRG): a score-orthogonal metric that asks whether a model's mistakes concentrate on the items humans found hard, or on items humans found easy. Together they expose, across a wide range of VLMs (and LLMs via OCR), three patterns: (i) low-budget accuracy collapses on the high-human-error tail at every model size; (ii) test-time scaling (TTS) raises token use roughly linearly with cohort error rate, while accuracy gains follow a non-monotonic curve; (iii) within a single family, TTS flips between anti-scaling on the hardest items and overthinking on easier ones -- two faces of the same alignment failure. On DRG, models with near-identical accuracy can sit at near-opposite values: one model gets wrong what humans also find hard, while another solves the hardest items yet fails on items humans find easy -- a contrast that aggregate accuracy hides. Our code and dataset builder will be open-sourced at https://github.com/naver-ai/KCSAT-ML.

Sanghee Park, Geewook Kim, Kee-Eung Kim

NAVER Cloud AI · KAIST AI

Mathematical Reasoning Benchmarks Large Reasoning Models Cohort

Jun 8, 2026cs.CL

SG-OPD: Sign-Gated On-Policy Distillation via Sign-Consistency Gating and Phased Teacher Sampling

On-policy distillation (OPD) trains a student on its own trajectories with dense per-token supervision from a stronger teacher, and often outperforms off-policy distillation and standard reinforcement learning. However, we find that its effectiveness implicitly relies on two assumptions that frequently break in practice: trajectory-level alignment between the student and the teacher, and uniform token-level reliability of the teacher's preferences. We therefore propose Sign-Gated On-Policy Distillation (SG-OPD), which uses a binary verifier as a trust signal for the teacher at two complementary granularities: phased teacher sampling mixes in verifier-endorsed teacher rollouts at cold-start, and a sign-consistency gate extrapolates the distillation update on tokens where the teacher agrees with the verifier-correct direction and interpolates it where it disagrees. Experiments on competition-level mathematical reasoning benchmarks show that SG-OPD consistently outperforms standard OPD, with average gains of 1.98 and 7.50 at the per-sample and per-question levels, respectively.

Haoran Xu, Hongyu Wang, Yifei Gao +3

Zhejiang University · Hunan University · Tianjin University +2

Efficient On-Policy Distillation Teacher Mathematical Reasoning Benchmarks

Jun 8, 2026cs.AI

A Regret Minimization Framework on Preference Learning in Large Language Models

Reinforcement learning with verifiable rewards (RLVR) has enabled progress on reasoning-intensive tasks by relying on task-specific verifiers that provide automated correctness signals. However, many realistic language tasks are difficult to equip with reliable verifiers, motivating a growing reliance on reinforcement learning from human feedback (RLHF). In this setting, we argue that a closer examination of how human feedback should be interpreted is essential. We introduce Regret-based Preference Optimization

(\textbf{RePO})

, which reframes RLHF through

\textit{regret minimization}

rather than reward maximization. Human preferences are often shaped by

\textit{prospective}

anticipation of outcomes and

\textit{counterfactual}

comparisons to alternative behaviors, rather than by immediate, outcome-independent utility.

\textbf{RePO}

captures this structure by modeling preferences as behavior-conditioned assessments of relative suboptimality. Experiments on mathematical reasoning benchmarks and human preference datasets demonstrate consistent performance gains, indicating that

\textbf{RePO}

is an effective and human-aligned approach for training large language models.

Suhwan Kim, Taehyun Cho, Geon-Hyeong Kim +4

Seoul National University, Seoul, South Korea · LG AI Research, Seoul, South Korea · UNIST, Ulsan, South Korea +1

Reinforcement Learning From Human Feedback Sequential Preference Optimization Preference Learning Verifiable Rewards Mathematical Reasoning Benchmarks

Jun 8, 2026cs.AI

Diverse Thinking Schemata Elicit Better Reasoning in Large Language Models

Large reasoning models (LRMs) have attracted increasing attention for their ability to solve complex mathematical problems by generating extended reasoning chains. In this work, we focus on two critical yet underexplored aspects of the reasoning process: reasoning transitions capturing the distinct transitions between reasoning steps and answer candidates reflecting the variety of solution paths produced by the model. We collectively define these two aspects as thinking schemata. We observe a correlation between the diversity of thinking schemata and model performance, which motivates us to enhance diversity as a means to further improve reasoning potential. To this end, we propose Diverse Schemata Policy Optimization (DiScO), a framework that first endows the model with schemata awareness, then encourages diversity through reinforcement learning, and further promotes diverse reasoning at inference time. Experiments on multiple mathematical reasoning benchmarks demonstrate that DiScO consistently outperforms standard group relative policy optimization. Beyond accuracy, human-annotated analyses show that DiScO substantially improves the model's ability to recover from erroneous initial attempts. Overall, our work suggests the important role that diversity of the thinking schemata plays and points to scaling along the diversity dimension as a promising research direction.

Xinyue Liang, Yizhe Yang, Yu Bai +3

School of Computer Science and Technology, Beijing Institute of Technology, Beijing, China

Large Reasoning Models Mathematical Reasoning Benchmarks Frictive Policy Optimization

Jun 3, 2026cs.LG

Selective-Advantage Entropy-Adaptive Horizon GRPO: Asymmetric Token-Level Discounting for Efficient Reinforcement Learning of Language Models

Group Relative Policy Optimisation (GRPO) has emerged as an effective reinforcement-learning algorithm for aligning language models on reasoning tasks, but it treats every token position and every sampled rollout symmetrically. We introduce two complementary extensions: (i) Adaptive-Horizon GRPO (AH-GRPO), which weights each token's policy gradient using a cumulative entropy-based discount that reduces the effective horizon when the model is uncertain, and (ii) Selective-Advantage AH-GRPO (SA-AH-GRPO), which applies this discounting only to negative-advantage rollouts, leaving positive-advantage, successful trajectories unattenuated. We evaluate standard GRPO with alpha = 0, AH-GRPO with alpha = 0.5, and SA-AH-GRPO with alpha = 0.5 on the GSM8K mathematical reasoning benchmark using both Qwen 2.5-1.5B-Instruct and Qwen 2.5-3B-Instruct fine-tuned with LoRA. On the 3B model, SA-AH-GRPO achieves Pass@1 = 0.858 at its peak at step 30 and maintains 0.846 at 180 steps, with training variance reduced to 0.0246, a 3.6 times reduction relative to GRPO while matching its peak accuracy. On the 1.5B model, SA-AH-GRPO achieves a peak Pass@1 of 0.686, improving over the zero-shot baseline of 0.637. Our analysis shows that asymmetric discounting preserves the full gradient signal on correct solutions, prevents entropy collapse, and substantially stabilises training, suggesting a principled inductive bias for reinforcement learning with verifiable rewards on structured generation tasks.

Chirag Chawla, Rohan Charudatt Salvi, Madhav S. Baidya

Indian Institute of Technology (BHU), Varanasi, India · Department of Computer Science, University of Illinois Chicago, IL 60607, USA

Policy Gradient Offline Reinforcement Learning Mathematical Reasoning Benchmarks Selective

Jun 3, 2026cs.CL

LoRi: Low-Rank Distillation for Implicit Reasoning

Implicit chain-of-thought (iCoT) methods aim to internalize reasoning in large language models, but often underperform explicit CoT prompting. We empirically find that hidden-state reasoning trajectories exhibit low-rank structure. Motivated by this observation, we propose a low-rank distillation framework that transfers reasoning by aligning teacher and student trajectories in a shared low-rank tensor subspace using first- and second-order statistics. The resulting formulation captures the global structure of reasoning while supporting a compact latent reasoning process. We evaluate the method across multiple model families, including LLaMA and Qwen, at different scales on mathematical reasoning benchmarks. Our approach consistently improves performance, especially on challenging multi-step tasks, approaching explicit CoT accuracy and outperforming prior iCoT distillation methods.

Ryan Solgi, Jiayi Tian, Zheng Zhang

University of California-Santa Barbara, USA

Chain-of-Thought Reasoning Reasoning Skills Mathematical Reasoning Benchmarks Dataset Distillation

Jun 2, 2026cs.CR

Testing LLM Arithmetic Reasoning Generalization with Automatic Numeric-Remapping Attacks

Large language models achieve strong performance on arithmetic reasoning benchmarks, and one common response to arithmetic brittleness is to delegate computation to code. Yet models are still often used in settings where they must reason directly from natural language, and trustworthy models should solve small-number arithmetic word problems without external tools. Prior work shows that LLMs are sensitive to numerical variation: a model may solve an original problem but fail on structurally similar variants requiring the same reasoning procedure with different numbers. We ask whether this fragility persists under a stricter setting involving small, schema-preserving numeric changes that retain the original reasoning program and avoid large-number stress tests. We introduce an automatic algorithm for generating numeric-remapping attacks on arithmetic word problems. Unlike template-based perturbation methods requiring manual schemas or constraints, our approach derives problem-specific symbolic representations, generates constrained numeric remappings, recomputes gold answers, and realizes transformed questions through deterministic edits guided by LLM-generated edit plans. Stage-wise validation and a high-confidence audit retain reliable attacks, making the pipeline scalable with limited human intervention. We evaluate DeepSeek-R1 (70B), Gemma4 (31B), and GPT-OSS (120B) on GSM8K, MAWPS, and MultiArith. On GSM8K, completed runs show conditional accuracy drops of 12.16 to 25.82 percentage points. MAWPS and MultiArith are far more stable, with most attacked accuracies near or above 98%. These results show that numeric-remapping robustness depends strongly on dataset structure: GSM8K remains sensitive even when reasoning programs are preserved and answers are recomputed, while shorter, more regular datasets are more robust.

Malia Barker, Bishal Lakha, Edoardo Serra +1

Department of Computer Science, Boise State University, Boise, ID, USA · University of L'Aquila, L'Aquila, Italy

Mathematical Reasoning Benchmarks LLM Reasoning Strategies Fragility Stress Testing Natural Language

Jun 2, 2026cs.LG

Right Makes Might: Aligning Verified Hidden States Empowers RL Reasoning

Reinforcement Learning from Verifiable Rewards (RLVR) has become the dominant approach for improving mathematical reasoning in large language models, yet current methods reduce each correct rollout to a single reward bit, ignoring the geometric structure shared among their hidden states. Investigating this structure, we find that at the anchor token (the position immediately before the answer marker), correct rollouts converge naturally because they must produce the same answer (cosine similarity ~0.84), yet each retains residual variance from its unique reasoning path. Encouraging full alignment at this point pushes the model to extract a unified "correct decision" representation, reducing sensitivity to which reasoning path was taken. Based on this observation, we propose Hidden-Align, an auxiliary loss function that aligns the last-layer hidden states of correct rollouts at the anchor token during RL training, with zero overhead in both training and inference. On eight mathematical reasoning benchmarks, Hidden-Align improves average pass@1 over the DAPO baseline by 3.8, 6.2, and 5.4 percentage points on Qwen3-1.7B, 4B, and 14B respectively, with consistent pass@k gains across all three scales, supported by ablations on loss type, anchor position, layer depth, and loss weight.

Ziyue Wang, Aomufei Yuan, Yongfu Zhu +10

1Peking University · 2JINGDONG · 3Shanghai Innovation Institute +2

Reinforcement Learning With Verifiable Reward Mathematical Reasoning Benchmarks Verifiable Rewards Value Alignment Stance

May 27, 2026cs.LG

FormInv: A Measurement Protocol for Semantic Invariance in Mathematical Reasoning Benchmarks

A paraphrase-quality audit of MathCheck (ICLR 2025) detected 4 semantically incorrect paraphrases in 129 groups (3.1%); removing them drops GPT-4o from rank 2 to rank 4 and elevates Claude Haiku and DeepSeek V3 above it; these ranking changes are invisible to any single-model evaluation. Cross-model unanimity found these errors automatically (>= 3/4 models for MathCheck; >= 6/9 for our primary evaluation) for under $10; in our own dataset the same protocol found that 47% of auto-generated connective-variation paraphrases were semantically incorrect. That flaw compounds a deeper measurement gap: Claude Haiku 4.5 achieves 86% accuracy yet SCR=50%, meaning half its theorems are answered differently under semantically equivalent restatements, while aggregate accuracy across 9 models spans only 86-96% yet Semantic Consistency Rates (SCR) span 50-82% -- a 32-point gap invisible to standard benchmarks. Formally, for any target ranking over 9 frontier models there exists a weighting over paraphrase families that realizes it (No-Free-Benchmark corollary), because no model Pareto-dominates all families -- so benchmark designers who select families are implicitly choosing which model wins. FormInv supplies the audit protocol (replicated on external benchmarks at 100% recall), SCR and per-theorem Cochran's Q as primary invariance measures evaluated on 9 models across 366-811 items (on Lean4-verified theorems), and FormInvSelector for regime-aware model selection.

Nishal Thomas, Noel Thomas

Independent Researcher · Mohamed Bin Zayed University of Artificial Intelligence

Mathematical Reasoning Benchmarks Theorem Proving Equivalence Model Auditing Measures

May 27, 2026cs.CL

Skill-Conditioned Gated Self-Distillation for LLM Reasoning

On-policy self-distillation (SD) improves LLM reasoning by using teacher-side privileged information (PI) to turn sparse verifier outcomes into dense token-level supervision. Existing methods usually assume trusted PI, such as reference answers or successful traces. We ask whether PI can instead come from an experience-derived skill bank, where retrieved skills are compact and reusable but may also be irrelevant or misleading. We propose Skill-Conditioned Gated Self-Distillation (SGSD), which formulates skill-based SD as teacher hypothesis validation rather than unconditional imitation. SGSD retrieves skill-mistake pairs, constructs a multi-teacher pool, and lets all skill-conditioned teachers score the same plain-prompt student rollout. The verifier validates each teacher's polarity: supporting a success or suppressing a failure gives positive supervision, while the opposite stance is reversed. A robust gated objective then distills informative teacher-student disagreements while suppressing uncertain or extreme signals. Experiments on multiple mathematical reasoning benchmarks show that SGSD consistently improves over GRPO and remains competitive with answer-conditioned OPSD under a weaker PI assumption. For example, on Qwen3-1.7B, SGSD outperforms GRPO by 6.2% and OPSD by 1.7% on average on AIME24, AIME25, and HMMT25.

Jiazhen Huang, Xiao Chen, Xiao Luo +3

Tsinghua University · University of Wisconsin-Madison · X-Humanoid +2

Unsupervised On-Policy Self-Distillation LLM Reasoning Strategies Mathematical Reasoning Benchmarks Teacher

May 27, 2026cs.LG

Interpretability-Guided Layer Selection over Subspace Projection: SAEs as Stethoscopes, Not Scalpels, for Raw Task Vector Model Editing

LLMs increasingly require surgical model editing to enhance domain-specific capabilities without incurring the computational cost or catastrophic forgetting associated with full fine-tuning. Sparse Autoencoders (SAEs) have emerged as a promising tool in this setting, in principle allowing for feature-level identification of where to intervene. In this work, we rigorously evaluate an SAE-guided editing pipeline for mathematical reasoning on Gemma-3-4B-IT and uncover a fundamental failure mode: the intuitively appealing approach of projecting task vectors onto SAE feature subspaces acts as an information bottleneck that discards approximately 97% of the modification energy, yielding no statistically significant improvements across seven math subjects. We show that this failure stems from a geometric misalignment between activation-space SAE directions and weight-space task vectors. We then propose a shift in perspective: SAE as a Stethoscope, Not a Scalpel, where SAEs are used for layer-level diagnosis rather than intervention-level filtering. By injecting unfiltered raw task vectors only into layers identified by an SAE-derived specificity score, we improve Number Theory accuracy from 29.6% to 39.4% (z=+3.41, p=0.0007) on the Minerva Math benchmark; 5 of 7 math subjects significantly improved and none significantly degraded. Our method is fully deterministic, requires no additional inference cost, and provides a principled framework for interpretability-guided model editing.

Li Lei, Madalina Ciobanu, Qingqing Mao +1

Incept Labs Houston, TX · Incept Labs, Houston, TX · Titan Holdings, San Francisco, CA

Mathematical Reasoning Benchmarks Subspace

May 27, 2026cs.LG

Joint Training of Multi-Token Prediction in Reinforcement Learning via Optimal Coefficient Calibration

Reinforcement Learning from Verifiable Rewards (RLVR) has emerged as the standard paradigm for improving reasoning capability of large language models, while Multi-Token Prediction (MTP) has been a widely adopted module in pretraining. Combining them is a natural approach, yet current RL practices detach MTP gradients because joint training degrades the performance. We revisit this failure from an optimization perspective. We show that the per-step effect of MTP on the RL objective can be decomposed into two terms: a first-order correlation and a second-order perturbation penalty. This decomposition unifies three MTP training regimes: Detach, Cross-Entropy loss, and Policy loss, and explains why each succeeds or fails. Further analysis of policy loss reveals that, although it aligns with intuition, performance still degrades: the correlation term decays while the quadratic penalty persists. Guided by the analysis, we propose Optimal Coefficient Calibration (OCC), an adaptive scheme that tracks the optimal coefficient online via a log-probability proxy at negligible cost. Across six competition-level mathematical reasoning benchmarks, OCC consistently matches or exceeds the detach baseline, delivering improved joint MTP-RL training performance.

Zili Wang, Jiajun Chai, Lin Chen +3

School of Artificial Intelligence, University of Chinese Academy of Sciences, China · MAIS, Institute of Automation, Chinese Academy of Sciences, China · Meituan

Reinforcement Learning With Verifiable Reward Multi-Token Prediction Mathematical Reasoning Benchmarks

May 26, 2026cs.CL

ReverseMath: Answer Inversion for Scalable and Verifiable Mathematical Problem Generation

Mathematical reasoning benchmarks are vital for evaluating large language models (LLMs), but many are static and repeatedly exposed through public evaluation and training pipelines, making it difficult to separate genuine reasoning from memorization. Meanwhile, manually constructing new math problems with reliable answers remains costly. We introduce ReverseMath, a scalable method for generating new math problems through answer inversion. Given a problem and its answer, ReverseMath masks a numerical value in the original problem, treats the original answer as a known condition, and rewrites the problem so that the masked value becomes the new answer. The generated problem reverses the original input-output relation, making its answer known by construction. We study ReverseMath for both evaluation and training. For evaluation, paired original/reversed problems reveal substantial behavioral shifts: models sometimes fail on reversed problems and even incorrectly output the original answer, suggesting memorization-like behavior. For training, ReverseMath provides automatically labeled reversed problems as data augmentation for reinforcement learning (RL). Experiments show that including ReverseMath-generated data improves mathematical reasoning performance across multiple benchmarks, demonstrating its value as both an analysis tool and a scalable source of verifiable training data.

Raoyuan Zhao, Yihong Liu, Yupei Du +2

Center for Information and Language Processing, LMU Munich · 2Munich Center for Machine Learning (MCML) · 3Saarland University

Mathematical Reasoning Benchmarks Reverse Memorization

May 26, 2026cs.AI

Reasoning, Code, or Both? How Large Language Models Handle Variations in Math Questions

Large Language Models (LLMs) achieve impressive accuracy on mathematical reasoning benchmarks, yet their performance drops when problems are modified with simple changes like different names or numbers. Code execution methods, which let models generate and run Python code instead of reasoning in natural language, have been proposed as a solution, but their effect on reasoning robustness (the ability to maintain accuracy across problem variations) has not been systematically tested. This study evaluates three approaches on 1,000 problems from the GSM-Symbolic dataset: pure reasoning using chain-of-thought (CoT) prompting, single-shot code execution using Program-Aided Language models (PAL), and iterative code execution using Step-by-Step Coding (SBSC). All three were run on paired original and modified problems using Claude Haiku 4.5. CoT was the most robust method, with an accuracy drop of 1.3 percentage points and 1.8% of problems breaking under perturbation. PAL was the least robust at 1.7 percentage points and 3.1% broke, with SBSC falling in between. Although these differences were not statistically significant (

p = .096

), the directional trend was consistent across all measures, suggesting that code execution, whether single-shot or iterative, does not improve reasoning robustness on grade-school-level problem variations.

Matthew Kutakh

Mathematical Reasoning Benchmarks LLM Reasoning Strategies Math Problems Chain-of-Thought Reasoning Language Modeling

May 21, 2026cs.LG

Tailoring Teaching to Aptitude: Direction-Adaptive Self-Distillation for LLM Reasoning

On-policy self-distillation (OPSD) is an emerging LLM post-training paradigm in which the model serves as its own teacher: conditioned on privileged information such as a reference trace or hint, the same policy provides dense token-level supervision on its own rollouts. However, recent studies show that OPSD degrades complex reasoning by suppressing predictive uncertainty, which supports exploration and hypothesis revision. Our token-level analysis shows that this failure arises from applying a uniform direction of teacher supervision across tokens with different uncertainty levels: conformity to the privileged self-teacher suppresses exploration at high entropy, while deviation from the teacher degrades step accuracy at low entropy. Accordingly, we propose \textbf{Direction-Adaptive Self-Distillation} (\textbf{DASD}), which reframes privileged self-distillation from uniform teacher imitation into entropy-routed directional supervision: high-entropy tokens are pushed away from the privileged teacher to preserve exploration, while low-entropy tokens are pulled toward the teacher to stabilize step-level execution. Across six mathematical reasoning benchmarks, DASD achieves the best macro Avg@16 over strong RLVR and self-distillation baselines. Pass@

k

, reasoning-health, and generalization analyses show that these average gains come from preserving exploration without sacrificing step-level execution.

Hongbin Zhang, Chaozheng Wang, Kehai Chen +4

Institute of Computing and Intelligence, Harbin Institute of Technology, Shenzhen, China · Keeta AI, Meituan, Beijing, China · Peng Cheng Laboratory, Shenzhen, China

Unsupervised On-Policy Self-Distillation Self-Teacher LLM Reasoning Strategies Mathematical Reasoning Benchmarks

May 21, 2026cs.LG

From Reasoning Chains to Verifiable Subproblems: Curriculum Reinforcement Learning Enables Credit Assignment for LLM Reasoning

Reinforcement learning from verifiable rewards (RLVR) has shown strong promise for LLM reasoning, but outcome-based RLVR remains inefficient on hard problems because correct final-answer rollouts are rare and sample-level credit assignment cannot use partial progress in failed attempts. We introduce SCRL (Subproblem Curriculum Reinforcement Learning), a curriculum RL framework that derives verifiable subproblems from reference reasoning chains and fixes the final subproblem as the original problem. This turns partial progress on hard problems into verifiable learning signals. Algorithmically, SCRL uses subproblem-level normalization, which normalizes rewards independently at each subproblem position and assigns the resulting advantages to the corresponding answer spans, enabling finer-grained credit assignment without external rubrics or reward models. Our analysis shows that subproblem curricula lift hard problems out of gradient dead zones, with larger relative gains as the original problem becomes harder. Across seven mathematical reasoning benchmarks, SCRL outperforms strong curriculum-learning baselines, improving average accuracy over GRPO by +4.1 points on Qwen3-4B-Base and +1.9 points on Qwen3-14B-Base. On AIME24, AIME25, and IMO-Bench, SCRL further improves pass@1 by +3.7 points and pass@64 by +4.6 points on Qwen3-4B-Base, indicating better exploration on hard reasoning problems.

Xitai Jiang, Zihan Tang, Wenze Lin +3

LeapLab, Tsinghua University · Qiuzhen College, Tsinghua University · B Gao Huang

Curriculum Reinforcement Learning Reinforcement Learning With Verifiable Reward LLM Reasoning Strategies Reasoning Chain Mathematical Reasoning Benchmarks Credit Assignment+1

May 20, 2026cs.LG

Advantage Collapse in Group Relative Policy Optimization: Diagnosis and Mitigation

Group Relative Policy Optimization (GRPO), a prominent algorithm within the Reinforcement Learning from Verifiable Rewards (RLVR) framework, has achieved strong results in improving the reasoning capabilities of large language models (LLMs). However, GRPO is prone to advantage collapse, a failure mode where homogeneous rewards within a group (e.g., all correct or all incorrect answers) yield near-zero advantages and vanishing gradients. To address this, we introduce the Advantage Collapse Rate (ACR), the first diagnostic metric quantifying the proportion of training batches with ineffective gradients. Across models from 0.5B to 14B parameters on mathematical reasoning benchmarks, we show that ACR strongly predicts training stagnation and final performance. We then propose Adaptive Virtual Sample Policy Optimization (AVSPO), a lightweight extension of GRPO that injects virtual reward samples, guided by real-time ACR monitoring, to enable learning from homogeneous groups without additional model rollouts. AVSPO reduces advantage collapse by 58-63% relative to GRPO and yields consistent accuracy gains of 4-6 percentage points across all model scales, while maintaining generalization on the evaluated out-of-domain task. Code and datasets are available at https://github.com/hexixiang/Advantage-Collapse-Rate.

Xixiang He, Qiyao Sun, Ao Cheng +5

National University of Defense Technology, Changsha, Hunan, China · Intelligent Game and Decision Lab, Beijing, China · The Chinese University of Hong Kong, Shenzhen, Guangdong, China

Reinforcement Learning With Verifiable Reward Collapse Mathematical Reasoning Benchmarks

May 19, 2026cs.AI

Beyond Fixed Budgets: Characterizing the Inelasticity and Limitations of Tree-of-Thought Reasoning Strategies

Tree of Thought (ToT) search has become a promising direction for improving the reasoning capabilities of large language models, but deploying these methods in practice raises a question that has received little systematic attention: how do different search strategies behave under varying compute budgets, model sizes, and problem difficulties? In this work, we evaluate two representative ToT methods; DPTS, a Monte Carlo tree search based approach, and SSDP, a semantic deduplication based approach, across two mathematical reasoning benchmarks (Math500 and GSM8K), two model scales (Llama-3B and Llama-8B), and four token budgets (3k--10k). Our analysis reveals that the two methods exhibit limitations that pull in opposite directions. DPTS suffers from a cold-start bottleneck at low budgets: it requires sufficient exploration before its value estimates become reliable, making it a poor fit for resource-constrained settings despite strong scaling behavior at higher budgets. SSDP, on the other hand, reaches candidate solutions efficiently but is prone to frontier depletion; its aggressive node merging permanently discards unexplored paths, leaving it unable to improve regardless of how much budget remains. Together, these findings suggest that neither a fixed exploration strategy nor a fixed pruning strategy is sufficient across compute continuum. We argue that effective search for scientific reasoning agents requires strategies that can adapt their behavior based on search progress and available resources.

Atkia Mahila, Avinash Maurya, M. Mustafa Rafique +1

Rochester Institute of Technology Rochester, USA · Argonne National Laboratory Lemont, USA

LLM Reasoning Strategies Monte Carlo Tree Search Mathematical Reasoning Benchmarks Thoughts Token Budget Allocation

May 19, 2026cs.LG

Detecting and Mitigating the Correct-Answer Extinction Window in Test-Time Reinforcement Learning with Majority Voting

Test-time reinforcement learning (TTRL) reports substantial accuracy gains on mathematical reasoning benchmarks using majority vote as a pseudo-label signal. We argue these gains are systematically misinterpreted: most reflect sharpening of already-solvable problems rather than genuine learning, while problems corrupted from correct to incorrect outnumber truly learned ones, and this damage is irreversible once majority vote locks onto a wrong answer. Per-problem tracking reveals that correct-answer signals in low-ability problems are briefly active before being permanently suppressed, a phenomenon we term the \textit{Correct-Answer Extinction Window}, with Flip Rate (FR) as its leading indicator. We thus propose TTRL-Guard, a lightweight framework with three mechanisms targeting the extinction window: Flip-Rate-Aware Reward Scaling (FRS) down-weights at-risk updates as FR declines, Minority-Preserving Sampling (MPS) retains gradient signal from minority correct answers, and Risk-Conditioned Sparse Updatings (RCSU) suspends updates on polarized problems. Experiments across three models and four benchmarks show that TTRL-Guard achieves the best average pass@1 on Qwen2.5-7B-Instruct and Qwen3-4B, improves relatively over TTRL by +54% on AIME 2025.

Hongxiang Lin, Zhirui Kuai, Erpeng Xue +1

Meituan

Reward Functions Mathematical Reasoning Benchmarks Majority Voting