Chain-Of-Thought Reasoning

Embodied agents using LLM-based planners often struggle with physical hallucinations, poor generalization to long-horizon tasks, and lack of environmental awareness. We propose GraphThink, a novel framework that integrates a task graph to provide structured knowledge for robust planning and a scene graph to maintain environmental memory for event-driven replanning. Specifically, the task graph guides LLM thinking through contextual prompting and iterative refinement, effectively mitigating planning hallucinations. Furthermore, within the GRPO framework, the task graph offers delicate reward design to train the LLM planner, enhancing long-horizon planning capabilities and improving generalization. Finally, an event-driven replanning module, powered by the scene graph, enables closed-loop environment awareness and error correction. GraphThink achieves state-of-the-art performance on the ALFRED benchmark. In particular, our high-level planner surpasses leading API-based LLMs on both the validation set and held-out long-horizon tasks, underscoring its robust zero-shot and few-shot capabilities. Additional evaluations further demonstrate strong out-of-distribution generalization to novel tasks and environments.

A host of recent studies have evaluated the cognitive and developmental alignment of Foundation Models (FMs). These investigations include evaluations of their correspondence to adult performance across a range of cognitive domains, as well as whether aspects of model training track children's cognitive development. However, using FMs as candidate cognitive models poses significant methodological and conceptual challenges. A key question underlies this effort: under what conditions does behavioral alignment justify treating FMs as explanatory models of cognition? In this paper, we articulate a four-stage inferential framework for evaluating FMs as cognitive and developmental models: adapting human experimental tasks to model-compatible formats, specifying linking hypotheses that map model outputs to human measures, evaluating behavioral correspondence, and comparing across candidate models or manipulations. We clarify the role of linking hypotheses in mapping model outputs to human behavioral measures, identify challenges that constrain alignment claims, and propose principles for theory-driven and comparative evaluation. Throughout, we argue that behavioral fit alone is insufficient. Alignment becomes scientifically meaningful only when embedded within explicit theoretical commitments, theory-diagnostic tasks, and systematic contrastive evaluation across candidate models.

Concept Bottleneck Models (CBMs) provide an interpretable framework by grounding predictions in human-understandable concepts, enabling semantic inspection and test-time intervention. Recent variants have improved CBMs through richer concept representations, uncertainty estimation, and dependency modeling. However, robust reasoning under unreliable concept states remains underexplored. Without such reasoning, misleading semantic evidence can propagate through the bottleneck, compromising both explanations and downstream predictions. To address this issue, we propose ReCBM, an uncertainty-gated relational reasoning framework for CBMs. ReCBM introduces semantically defined concept relations into the bottleneck and uses uncertainty to guide their refinement. By modeling co-occurrence, implication, and exclusion, ReCBM specifies how evidence is exchanged across concepts, while uncertainty modulates the contribution of each concept during this process. Experiments across diverse datasets showed that ReCBM improved concept and task recovery under missing and flipped concepts, supported uncertainty-aware intervention, and extracted compact task-relevant concept subsets without degrading downstream performance.

Test-time scaling is often implemented by spending more compute along one axis: sampling more solutions, extending a chain of thought, or applying a stronger evaluator. Under a fixed inference budget, these choices compete. This paper formulates test-time reasoning as a compute-allocation problem in which a system must decide whether the next unit of compute should be spent on generation, verification, or stopping. We introduce CoBa, a compute-balanced routing policy that first obtains a small set of candidates, applies cheap verification broadly, and routes uncertain or high-value candidates to stronger verification. On 3,129 example-generator evaluations spanning MATH-500, AIME 2024/2025, AMC 2023, and procedural symbolic reasoning, CoBa-Routed-Strong reaches 85.13% macro accuracy, statistically matching a self-evaluation weighted-voting proxy at 85.20% while using 49.1% fewer parameter-weighted tokens. It also matches best-of-16 majority voting within 0.01 macro-accuracy points while using 58.9% fewer parameter-weighted tokens; paired tests retain a small best-of-16 edge at substantially higher cost. Paired bootstrap tests show significant gains over single-sample decoding, while the remaining gap to the pool oracle exposes headroom for sharper routing. For local reasoning systems, test-time scaling becomes a question of where the next computation is most valuable.

The Tower of Hanoi is a simple planning puzzle that in prior work has proven challenging for large reasoning models (LRMs). Current models solve the standard formulation of the puzzle, but still struggle with the flat-to-flat variant (where initial and goal states are not restricted to have all rings on a single peg). This paper presents an in-depth study of how both small, in-house Transformers and large, third-party LRMs solve this task. To understand the failures mechanistically, we first train small Transformers from scratch on precomputed solution traces. Using a variety of interpretability techniques, we show that these Transformers develop an emergent world model: a linearly decodable, geometrically faithful representation of the puzzle's state space (the Sierpinski triangle), that is causally involved in solving the puzzles. Second, we return to the large LLMs and apply our techniques to two frontier reasoning models, Qwen3.6-27B and DeepSeek-R1-Distill-Qwen-32B, that attempt to solve the task through extended chain-of-thought. Surprisingly, we find that both models encode the Sierpinski world model near-perfectly at the end of the prompt, and yet fail at the majority of tasks when there are more than 3 rings. We locate the source of this failure in the decaying representation of the world model. We probe for the representation at different stages during planning, and establish causality by showing that performance can be improved by injecting the prompt-time representation at inference. The failure of the models is thus one of maintenance of the required representations, not their absence, and performance is at least partially recoverable. These results thus reframe the reported collapse in performance from prior work: current Large Reasoning Models build a world model, and then lose it.

World Action Models (WAMs) aim to construct a unified architecture capable of understanding world state evolution and guiding to generative motion planning. However, existing visual branches focus on predicting static visual observation, rather than reflecting potential transition information that captures the evolution of world states under motion interactions. This leads to representational entanglement between high-level physical condition evolution and low-level action trajectory generation within the Action Model, creating a structural bottleneck while weakening the predictive capability of world evolution modeling for action generation. We propose PILOT (Physical Inference for Latent Optimized Trajectories), whose core Representational Deduction (RD) bridges this gap by integrating motion thought-of-chain (CoT) guidance as a native model capability. Specifically, RD aims to encourage the action branch to explicitly model potential state transition tokens, which are retained as CoT in the reasoning space to guide fine-grained motion trajectory. Experiments demonstrate that RD not only significantly improves the success rate and generalization ability of WAMs in complex robotic manipulation tasks but also enhances the model's physical interpretability by decoupling high-level motion semantics from low-level trajectory details. Furthermore, the abundant state transition supervision signals introduced by RD effectively alleviate the sparse supervision in action generation, enabling it to serve as an efficient few-shot real-robot fine-tuning strategy and demonstrating superior scalability for migration to mainstream WAM architectures.

Vision-Language-Action (VLA) models have recently achieved impressive performance for end-to-end autonomous driving, yet existing approaches are primarily designed for an individual single autonomous driving agent with limited support for cooperative perception, reasoning, and planning. We present Cooperative Multi-agent Unified Driving with Reasoning (CMU-Drive), a closed-loop end-to-end benchmark for evaluating cooperative autonomous driving with multiple connected autonomous vehicles (CAVs) operating in safety-critical driving scenarios with background traffic participants. We further propose Vehicle-to-Vehicle Vision-Language-Action (V2V-VLA), a cooperative VLA model that integrates cooperative driving into a single forward pass by jointly generating driving actions, future waypoints, language reasoning, and communication policies. Experiments on CMU-Drive establish the first benchmark and baseline for cooperative VLA driving and provide a foundation for future research on multi-agent, closed-loop, end-to-end cooperative autonomous driving. Our code, benchmark, and model checkpoint will be publicly released to facilitate open-source research.

Token-level collaboration allows a large language model (LLM) to assist a small language model (SLM) when their predictions diverge. Existing methods either use LLM-generated intervention tokens or rank candidates with the LLM's next-token probabilities. Both rely on the LLM's local preference, even though an LLM-selected token may be difficult for the SLM to build on. We present FutureBridge, which ranks joint LLM-SLM token candidates according to how well they support the SLM's subsequent reasoning. During training, an answer-verified LLM trajectory supplies a fixed shared future, and a frozen SLM evaluates every candidate under this common context. The resulting counterfactual scores supervise a lightweight token reranker that observes only the current state and candidate token. At inference, FutureBridge uses the LLM only to expand the candidate pool, selects one token, and returns generation to the SLM without generating or appending a future suffix. Across five mathematical reasoning benchmarks, FutureBridge improves the Qwen3-1.7B SLM's Math Avg. by 35.1% relative to greedy SLM decoding. These results indicate that token selection benefits from modeling whether the receiving SLM can use each candidate to continue reasoning, rather than relying on the LLM's local preference alone.

Reasoning in large language models (LLMs) is often grounded in human text, human demonstrations, and human-generated rationales. For equilibrium reasoning in complex games, however, relying on human data can be suboptimal. In fact, human play is often guided by intuition and heuristics and can deviate substantially from game equilibrium. This discrepancy is amplified in games with mixed-strategy equilibria, where human data is heavily biased toward pure strategies. Consequently, conditioning LLMs on this data yields weak game strategies. To grant LLMs the reasoning capacity in games, in this work, we study how to elicit equilibrium play using solver output. We propose Mixed-Strategy Decision Tree (MDT), which articulates the silent optimality of the equilibrium into sparse strategic rules that both humans and LLMs could understand. Using solver output rather than human annotation allows us to extend the input to arbitrarily new states and continuations. We instantiate this study on No-Limit Texas Hold'em by querying a solver oracle for over \textbf{250 million mixed-strategy decisions}; MDT together with other techniques \textbf{reduces the $\ell_1$ distance to the equilibrium by $52.6\%$} across $8$ different LLM configurations. A Route-only ablation tests the incremental contribution of the shadow-based contrast, while complete River-endgame and Liar's Dice experiments evaluate strategic fidelity and portability beyond the original NLH communication setting.

Vision-language-action (VLA) models provide strong semantic priors for robot navigation, but they often ignore embodiment-specific mobility constraints. A path that is semantically plausible for one robot may be physically infeasible for another. We propose CrossTracer, a hierarchical framework for cross-embodiment navigation through adaptive trace residuals. CrossTracer represents navigation plans as normalized image-plane waypoints, forming a unified pixel-space interface between semantic reasoning and physical grounding. First, Vision-Language Trace Proposer (VL-Tracer) adapts a pretrained VLA model to predict an initial navigation trace from egocentric observations and flexible goal specifications. Second, CE-Adapter refines this trace by predicting embodiment-conditioned residual corrections from visual traversability cues, robot identity, and the initial trace. To train the refinement module without costly manual annotation, Cross-Embodiment RRT* (CE-RRT*) converts panoptic segmentation into robot-conditioned traversability cost maps and generates cost-minimizing pixel-space traces. We evaluate CrossTracer on the NaviTrace benchmark, which tests whether a model can generate embodiment-consistent navigation traces from egocentric observations, language instructions, and robot embodiment types. CrossTracer achieves a total score of 45.68, outperforming the strongest evaluated general-purpose baseline, Gemini-2.5-Pro, by 10.01 points, corresponding to a 28.1% relative improvement. Real-world deployment on wheeled and legged robots further shows improved navigation success and execution efficiency.

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.

Unified multimodal retrieval aims to identify candidates that satisfy complex user intent expressed through heterogeneous inputs. Although Large Vision-Language Model (LVLM)-based retrievers are efficient and scalable, directly encoding raw multimodal inputs often misses fine-grained discriminative cues, leading to confusion among semantically similar candidates. Recent methods mitigate this limitation by generating Chain-of-Thought (CoT) rationales to enrich the query representation. However, such reasoning is typically derived from the query alone: it explains what the query describes, but not what the retriever misunderstands. We argue that effective retrieval reasoning should instead be conditioned on retrieval feedback. Based on this insight, we introduce UniME-R1, an embedder-adviser framework that learns to reason over initially retrieved candidates and generate Retrieval-Centric Chain-of-Thought (RC-CoT). The adviser analyzes candidates individually to identify the discriminative cues confused by the embedder. If the target appears in the initial top-k set, UniME-R1 directly reranks the candidates; otherwise, it generates RC-CoT to refine the retrieval direction and performs full-corpus re-retrieval with a dual-mode embedder. To train the framework, we mine hard negatives to simulate realistic retrieval failures, jointly optimize direct retrieval and RC-CoT-augmented retrieval, and align the adviser with retrieval outcomes through supervised learning and retrieval-oriented reinforcement learning. Extensive experiments on MMEB-V2 and a diverse set of general multimodal retrieval benchmarks demonstrate that UniME-R1 consistently improves retrieval performance over strong baselines.

On-Policy Distillation (OPD) is emerging as a promising alternative to reinforcement learning for LLM post-training, yet its effectiveness in multilingual settings remains underexplored. We study OPD and its advanced variant, On-Policy Delta Distillation (OPD$^2$), for mathematical reasoning in English, Korean, and Japanese. OPD$^2$ improves OPD by using the probability gap between a post-trained teacher and its base model as the learning signal. Experiments with Qwen3 show that OPD$^2$ consistently outperforms the original OPD, with particularly strong improvements in Korean and Japanese, and generally narrows the English-Korean performance gap. We further find that English-only OPD can also increase performance for Korean and Japanese, but often shifts the responses toward English, highlighting the importance of multilingual data to preserving target-language responses.

Masked diffusion language models (dLLMs) can commit tokens in any order -- a freedom marketed as their core advantage over autoregressive decoding. We show that on reasoning tasks this freedom is instead the axis of failure. Logging every commitment during decoding of LLaDA-8B on GSM8K, we find that unconstrained (pure) decoding commits the final answer at 15-24% of the trajectory while half the reasoning region is still masked, and collapses to answer-only outputs on up to 90% of problems as the canvas grows. The cause is not the model's termination beliefs -- EOS "pressure" is nearly identical across decoders -- but reachability: whether the sampler may act on those beliefs at distant positions. A 2x2 prompt-decoder design shows that chain-of-thought helps only under ordered commitment (interaction +34.8 percentage points, 95% CI [26.8, 42.8]; without reasoning text the decoders are indistinguishable), an interaction we decompose into a collapse channel and an order channel and replicate on Dream-7B and MATH-500. A single-knob intervention -- frontier-gated commitment -- causally recovers the full gap (0.528 to 0.852) while preserving up to 4x parallel decoding, along a measured frontier whose optimal window flips from w=1 at full refinement to unconstrained at 8 tokens/step. Our results reframe existing window-style samplers, previously motivated by efficiency, as the minimal fix for a reasoning pathology they were never designed to address.

As language models are increasingly used for tasks that require verifiable reasoning, reliably distinguishing sound reasoning from flawed reasoning has become an important practical problem. Recent trajectory-based methods seek this signal in layerwise residual-stream displacements, which capture how representations change while attenuating some stable, token-specific information. However, displacement omits the state from which an update originates, whereas restoring the full state risks reintroducing shortcut-prone information. We identify this trade-off and propose a three-stream detector that combines motion with two restricted views of location. A coarse region reader based on vector quantization and a fine direction reader over normalized multi-layer states. This design restores enough state context to interpret the motion without returning to full-state probing. On reasoning benchmarks unseen during training, our method improves selection accuracy by up to 12% over the displacement-only state of the art and 21% over single-layer probing baselines. Although trained only on reasoning benchmarks, it also reads factual completion and fact verification, ahead of every detector we compare against, which places the signal on correctness rather than on a kind of reasoning. Ablations further show that motion, region, and direction provide complementary signals. These results suggest that reasoning validity is better read from state-conditioned motion than from either static states or decontextualized trajectories alone.

Test-time scaling improves LLM reasoning by using additional inference compute, but wider sampling alone can suffer from diminishing returns: new rollouts often repeat existing answer patterns instead of adding useful reasoning diversity. Verifier-based selection offers an alternative, but its performance depends on the calibration of an external reward model. We propose a verifier-free breadth--depth refinement framework that uses test-time compute to both explore and improve candidate solutions. The method samples multiple independent reasoning rollouts, refines each rollout through iterative self-critique and self-correction, and aggregates the refined answers by majority voting. Breadth preserves diverse initial attempts, while depth repairs local reasoning errors before aggregation. Across AIME24, AIME25, AMC, OlympiadBench, and MATH500, our method consistently improves over greedy decoding, majority voting, verifier-based best-of-$N$, beam search, and lookahead decoding across multiple open-weight models. For instance, with Qwen2.5-1.5B, accuracy increases from the strongest verifier-based baseline to $58.0\%$ on MATH500, and from $25.0\%$ to $32.5\%$ on AMC. These results show that test-time compute can be more effective when used to refine sampled trajectories rather than only to sample more candidates or rely on verifier-guided selection.

Video object removal must eliminate not only the target object but also its induced effects while maintaining high-fidelity and spatiotemporally coherent restoration. Existing methods mainly learn object-effect correspondences implicitly from predefined effect categories and fixed data distributions, limiting their generalization to complex real-world scenes involving compositional effects, spatially detached or weakly correlated effects, long-tail physical phenomena, and dynamically evolving interactions. We propose EffectLearner, a semantic-reasoning-enhanced framework that combines a VLM-based Object-Effect Reasoner with a DiT-based Video Eraser. Guided by a structured effect-analysis prompt, the Reasoner performs cross-modal reasoning over a target-highlighted video and extracts compact effect-aware context, which guides the Video Eraser toward comprehensive object-effect removal. Motion-aware mask guidance and motion-consistency supervision further improve removal coverage and spatiotemporal stability under object motion and evolving scene dynamics. To fully exploit the framework in challenging real-world scenarios, we further construct EffectWorld, a paired video dataset specifically designed for complex object-induced effects, and introduce a progressive training curriculum that combines common supervision with complex-effect data. On the standard ROSE-Bench, EffectLearner outperforms existing baselines on most metrics and achieves clear advantages on both EffectWorld-Eval and the challenging EffectWorld-Wild, demonstrating its ability to deliver high-quality video object removal in complex real-world scenes.

Evolution Strategy (ES) is a promising alternative to gradient-based fine-tuning for resource-constrained Large Language Model (LLM) reasoning. However, directly applying ES to billion-parameter LLMs is highly ineffective. In such high-dimensional parameter spaces, most random perturbations are nearly orthogonal to useful update directions, leading to unstable optimization. We propose Hyper-ES, a subspace-based ES framework that avoids the weakness of ES in full-parameter search while exploiting its strength in low-dimensional optimization. Instead of asking ES to discover useful directions from random perturbations in the LLM parameter space, Hyper-ES first performs a small number of inexpensive gradient-based fine-tuning runs to obtain descent directions. Although each direction may provide only a limited improvement on its own, their span forms a compact adaptation subspace that captures useful reasoning updates. Hyper-ES then applies CMA-ES to optimize layer-wise DARE-TIES merging coefficients within this subspace, allowing ES to search over combinations of meaningful descent directions rather than over arbitrary full-model perturbations. We evaluate Hyper-ES on three Qwen2.5-Instruct and DeepSeek-R1-Distill backbones across six mathematical reasoning datasets. Results show that Hyper-ES consistently outperforms GRPO-LoRA by 1% while requiring 10% fewer space-consuming gradient updates. Code at https://github.com/kuangrepi/Hyper-ES.

Interpretability methods for large language models (LLMs) describe internal state but do not directly test whether that state is causally sufficient to produce the observed behavior. In earlier work, we lesioned LLMs to produce error profiles in picture naming, a central task for assessing aphasia, and found that specific lesions produced errors resembling those of individual stroke survivors. Here we ask the inverse question: given an error profile, can the lesion parameters that produced it be recovered, and what does this inverse problem reveal about transformer computation? Lesions in LLaVA-Vicuna 13B were parameterized by layer index, modification percentage, and noise sigma across 4,840 configurations, and error profiles were characterized by a seven-category clinical taxonomy (correct, semantic, unrelated, formal, mixed, neologism, no-response). We trained a multi-task neural network to map error profiles back to perturbation parameters. The problem admitted a partial solution: across 10 independently trained inverse models, modification percentage and noise sigma were recoverable, whereas layer index was recoverable only within a neighborhood. In counterfactual validation, a fresh model instance perturbed with the recovered parameters reproduced the target behavior in 81.4% of cases. This dissociation between low layer recovery and high counterfactual fidelity is consistent with functional redundancy across transformer layers, a property not captured by standard interpretability methods. As an out-of-distribution test, we applied the trained model to picture-naming error profiles from 278 stroke survivors; recovered parameters were syndrome-discriminative, most strongly for perturbation intensity, indicating generalization beyond the training distribution. Counterfactual validation provides a general framework for LLM interpretability claims beyond inverse mapping.

Long context reasoning in large language models (LLMs) is usually constrained by the fact that a single inference trajectory has to simultaneously explore the context, store intermediate state, verify evidence, and produce the final answer. This becomes particularly difficult in tasks that require extraction, counting, ordering, or multi-hop reasoning, where an early mistake can propagate until the final response. In this work, we propose Chained Recursive Language Models (Chained RLM), an inference-time architecture, in which the same underlying model is called repeatedly as a sequence of fresh reasoning roots. Each root receives the original problem and context, but does not inherit the full conversational history. Instead, it receives a compact plain-text summary, a plain-text blackboard, and some durable task-specific artifacts written by predecessor roots. The motivation is to manage the context by chopping into partial tasks rather than one large inference response; in each staged computation, intermediate artifacts can be inspected, corrected, and extended by a later fresh inference by the same model. We describe the system model, handoff mechanism, artifact workspace, and evaluation protocol for this system. We study when fresh-context artifact continuation gives a measurable gain in accuracy over direct LLM answering even with recursive tool-calling.

Chain-of-thought (CoT) reasoning offers a window into the decision-making of large language models (LLMs), which can be monitored for target behaviors by reading the reasoning trace, motivating work on CoT monitorability. Latent CoT approaches, however, replace the explicit tokens with a small number of continuous states, lowering inference costs but removing the readable trace this monitoring relies on. Monitoring then requires alternative access to the model, such as probing its activations or verbalizing the latent states back into text, but how much monitorability these alternatives preserve is unclear. We study this question with a hint-based intervention setup, a proxy for behaviors where models exploit biasing input cues, e.g., an inadvertently leaked answer or a belief stated by the user, without acknowledging them. Taking hint-reliance as the monitorability target, we compare monitors across reasoning modes, from explicit CoT to weakly- and strongly-supervised latent CoT, on math reasoning and question answering. We find that, in this setup, monitorability depends more on properties of the task (such as whether the correct answer constrains the supporting reasoning) and the level of access to model internals than on the reasoning mode.

Research agents increasingly conduct multi-round machine-learning experiments in industrial recommendation settings and retain the resulting trajectories to guide later decisions. Yet a completed trajectory is not automatically evidence: generated artifacts may be unsupported or incomplete, executed rounds may be invalid or confounded, and later modifications may obscure earlier findings. We study \textbf{trajectory-to-evidence conversion}, asking what a completed research process has actually established. We introduce an evidence-grounded framework that couples bounded verification of consequential artifacts with post-execution claim qualification. A context-isolated generate--verify--repair process checks artifacts for evidence violations and missing downstream requirements before release. After execution, validity and attribution checks consolidate evidence across rounds, qualify intervention-level claims as actionable repairs, diagnostic guards, or withheld findings, and preserve admitted claims as auditable records with explicit provenance and applicability boundaries. A hybrid LLM-assisted controller subsequently applies, defers, or rejects records based on available target evidence. Record audits characterize which claims survive qualification, while downstream diagnostics identify affirmative applicability judgment as a bottleneck for the tested controller. Across paper-to-target adaptations, later rounds often improve on the first, while final rounds frequently underperform an earlier best, exposing non-monotonic trajectory evolution. Candidates produced through the complete workflow also yielded positive online lifts relative to deployed baselines.

Large Reasoning Models (LRMs) excel on complex tasks through long chain-of-thought (CoT) reasoning, but their lengthy intermediate steps cause severe overthinking that inflates inference cost. KV-cache compression is a common solution, yet existing reasoning-oriented methods apply a uniform policy across the trajectory and judge compression only by what it removes from the cache. Two observations point the other way. First, a reasoning state's tolerance to context loss varies along the trajectory, and process reward tracks it: deleting tokens at high-reward steps preserves accuracy far better than deleting the same budget at random. Second, compression is not free on the generation side, since a smaller cache leads the model to generate more tokens, partly canceling the saving. Together these motivate coordinating both sides under a single process reward. We propose ReCo (Reward-Coordinated Compression), a step-wise framework in which a lightweight process-reward estimator scores each completed step and drives three components: (1) reward-adaptive KV-cache compression that shrinks the retained cache harder at high-reward steps and less at low-reward ones, (2) a reward-banded penalty on reflection tokens that curbs redundant generation, and (3) confidence-based early stopping that triggers when the reasoning is reliable. Across three reasoning models and six benchmarks, ReCo reduces generated tokens by 37%-65% and end-to-end latency by 2.08x-2.35x over Full CoT, all while largely preserving accuracy.

Although Multimodal Large Language Models (MLLMs) have made substantial progress, their spatial reasoning may still produce intermediate judgments inconsistent with the input image, allowing errors to propagate through the reasoning chain and affect the final answer. Existing methods mainly improve spatial reasoning through training or additional spatial information, without considering whether the reasoning process itself is faithful to the model input. Our study shows that unfaithful reasoning chains significantly reduce final-answer accuracy. To address this issue, we propose a modular and training-free framework for spatial reasoning verification and correction. The framework constructs a Spatial Evidence Graph (SEG), which associates atomic spatial evidence extracted from Chain-of-Thought reasoning with visual entities, spatial relations, source steps, and visual evidence. Spatial Evidence Reliability Assessment (SERA) evaluates the reliability of visual evidence based on object existence, localization, and geometric measurements. The framework then identifies the earliest spatial evidence unit contradicted by reliable visual evidence and guides the original MLLM to revise the subsequent reasoning and final answer. Across 15 model-dataset settings, our method achieves an average accuracy of 68.94%, outperforming the compared baselines by 8.55 percentage points on average. Our code will be open-sourced.

Chain-of-thought (CoT) monitoring is increasingly treated as an important safety layer for frontier reasoning models. Most monitorability evaluations study explicit-influence settings: setups where the prompt directly incentivizes the model to hide something, e.g., by instructing it to perform a hidden side-task. A complementary axis for CoT-monitor evaluations is implicit-influence settings, where the prompt contains no instruction to hide, but the model's behavior is still shaped by features of the task or context, e.g. an irrelevant detail about a candidate that biases a hiring rating. We introduce the first benchmark that directly compares CoT monitorability under the two regimes. We test how model behavior changes in the presence of a nudge to choose a particular option. The nudge is delivered either as a casual aside (implicit), or as a direct instruction to act on the nudge and to conceal having done so (explicit). The benchmark spans four task formats (binary choice, numeric rating, multiple-choice QA, open-ended coding) and seven frontier extended-thinking models. Under explicit influence, a CoT monitor detects 60-94% of behavior shifts: even models instructed to conceal it leak the instruction into their CoT. Under implicit influence, the same factors still shift behavior, but detection falls by 41-46 percentage points in two of our four settings. Realistic system-prompt additions (of the kind a developer might deploy to reduce off-topic bias) lower implicit detection further, to as low as 5%, while preserving the behavioral influence itself. These results suggest that monitorability estimates obtained in explicit-influence settings may over-estimate monitorability, and that monitorability can be further decreased by well-intentioned deployment choices. Our benchmark and code are available at https://github.com/agatha-duzan/implicit-vs-explicit-influence

Large Language Models (LLMs) have transformed computational linguistics and achieved remarkable performance across numerous natural language processing tasks, yet significant gaps persist in understanding how these systems process culturally embedded linguistic expressions. This paper introduces ProverbIT, a novel Italian benchmark comprising 100 multiple-choice questions designed to evaluate LLMs' ability to complete Italian proverbs. We assess 13 frontier models, including Large Reasoning Models (LRMs) and traditional LLMs, across three tasks: proverb completion, multiple-choice selection with correct answers, and multiple-choice selection without correct answers. Our evaluation reveals surprising results: while nearly all models demonstrate knowledge of the proverbs through successful completion tasks, performance drops dramatically when transitioning to multiple-choice formats without correct answers, with even state-of-the-art reasoning models showing substantial degradation. Through detailed Chain-of-Thought analysis of two LRMs, we uncover that models exhibit a strong bias toward selecting literal synonyms and frequently mention correct proverb endings during reasoning without successfully identifying their absence from the given options. These findings suggest that current LLMs rely heavily on memorized patterns rather than deeper semantic understanding of culturally grounded expressions, highlighting important limitations in their reasoning capabilities for figurative language comprehension.

Large language models (LLMs) have recently shown strong performance on Theory of Mind (ToM) tests, prompting debate about the nature and validity of the underlying capabilities. At the same time, reasoning-oriented LLMs trained via reinforcement learning with verifiable rewards have demonstrated notable improvements across a range of benchmarks. In this work, we examine the behavior of such reasoning models in ToM tasks using novel adaptations of machine psychological experiments together with results from established benchmarks. We observe that reasoning models consistently exhibit increased robustness to prompt variations and task perturbations. Our analysis suggests these gains come at least partly from models being more robust at reaching the correct answer under prompt and task variation. We read this as evidence for a robustness-based account rather than for a new ToM-specific ability.

LLM-based multi-agent systems (MAS) typically optimize a single topology, restricting reasoning to a narrow trajectory and limiting comprehensive analytical capacity. Naively merging multiple topologies into a composite graph introduces redundant noise propagation across irrelevant connections, degrading solution quality. To address this dilemma, we propose \textbf{Hierarchical Sparse Coordination over a Union of Complementary Topologies for MAS (HELENA)}, a multi-agent framework that balances diverse reasoning paths with sparse task-dependent execution. \helena{} constructs a union MAS graph from complementary candidate topologies selected via Monte Carlo Tree Search and Determinantal Point Process, broadening the reasoning trajectory for comprehensive analysis of complex problems. A Hierarchical Sparse Coordination module then activates only a sparse subgraph at each step while agents exchange compressed latent briefs to suppress redundant noise propagation. Finally, a Local Self-Refinement stage identifies decision units with discrepancy evidence and rewrites them only when contrastive evidence simultaneously confirms a reliable solution-side failure and a challenger-side improvement. Experiments across eight benchmarks show that \helena{} achieves state-of-the-art results on all benchmarks, with an average gain of \pctup{3.47} over the strongest baseline and up to \pctup{10.34} on MMLU-Pro, achieving larger improvements on harder benchmarks at a reasonable additional cost.

Synthetic generation of Cognitive Behavioral Therapy (CBT) sessions is challenged by two competing demands: adhering to strict therapeutic structure while modeling the resistant, unpredictable behavior of real patients. Existing script-based methods fail to capture dynamic therapeutic interactions, while multi-agent approaches struggle to adhere to CBT's sequential structure; both suffer from sycophancy, producing overly compliant patients that misrepresent real clinical settings. In this work we introduce ODRA, a novel framework for synthesizing therapy dialogues through a Chain-of-Thought (CoT) strategy grounded in foundational CBT guidelines (Beck, 2020). ODRA further incorporates a resistance orchestrator to solve patient sycophancy, which employs steering techniques to elicit behaviors aligned with their resistance level. Automated and expert evaluations show that ODRA significantly outperforms existing methods across therapeutic skills, CBT alignment, and patient behavioral fidelity, with licensed psychologists preferring ODRA sessions across 12 of 13 clinical metrics. Furthermore, models fine-tuned on our dataset demonstrate superior therapeutic performance against both cooperative and resistant patients, validating that explicit resistance modeling in synthetic training data directly translates to downstream clinical robustness.

Existing document understanding benchmarks have largely focused on locating page elements, yet real-world document intelligence requires models to reason jointly about region semantics, spatial relations, and visual structure. We present ADOPD 2026, a reasoning-oriented extension of ADOPD that turns page decomposition into spatially grounded document understanding. ADOPD 2026 enriches page anchors inherited from ADOPD 2024 dataset with human-cleaned captions, semantic tags, and generated chain-of-thought (CoT) traces grounded to document regions. Instead of treating boxes, masks, and tags as independent supervision signals, we cast text blocks, visual entities, semantic labels, bounding boxes, and polygon masks as a shared vocabulary of visual anchors. This representation supports three connected capabilities. First, region-level semantic tagging asks models to identify document element types from both page context and local appearance, revealing long-tail semantic failures that standard layout benchmarks often hide. Second, unified vision-language grounding generates text regions and visual entities together with coordinates or polygonal outlines, transforming detection and segmentation outputs into structured anchors that can be reused by downstream reasoning systems. Third, current state-of-the-art models still struggle with dense counting tasks evaluated on DocCount, a benchmark derived from ADOPD 2026, highlighting the need for the Thinking-with-Anchors pipeline in document semantic understanding. By connecting page decomposition to verifiable visual-anchor reasoning, ADOPD 2026 provides a task framework that moves document understanding beyond localization toward anchor-grounded document intelligence.

We present a structured large-language-model (LLM) architecture for zero-shot human--robot coordination in a cooperative construction task with private goal views. Guided by a Dec-POMDP formulation, the architecture decomposes decision-making into (i) action-conditioned Theory-of-Mind (ToM) inference, (ii) hierarchical planning, (iii) conversation interpretation, (iv) action verification, and (v) feedback-based replanning. We compare the proposed method with an ablation without ToM inference and a multi-agent reinforcement-learning policy trained offline over many goal pairs. In human-participant experiments, the proposed method required fewer interaction steps and yielded higher post-interaction trust ratings than both baselines. These results suggest that systematically decomposing the team decision problem, using LLMs as tractable surrogates for otherwise intractable inference and planning computations, and retaining conventional verification for physical feasibility can improve both task coordination and the human experience.

Large language models can solve harder reasoning problems with more inference-time compute. The term "test-time scaling," however, covers several inference algorithms: extending deliberation along one trajectory, sampling completed candidates and aggregating them by voting or verification, and searching over partial states. These algorithms differ in statistical structure, compute requirements, and failure modes. Treating them as interchangeable under a scalar "budget," or reporting accuracy without specifying the inference protocol, makes results difficult to compare across studies. We study test-time scaling along three axes. First, we formalize it as budgeted inference over the implicit prefix tree of an autoregressive model and distinguish single-trajectory sequential scaling, leaf-level scaling with terminal reduction, and prefix-level scaling. Second, we treat the full inference system as the evaluated object and separate end-to-end performance from candidate-bank diagnostics. We introduce an evaluation profile whose coordinates and simple functionals recover or bound common repeated-sampling metrics, and require compute accounting and uncertainty estimates that match the protocol. Third, we distinguish exact replay from distributional reproducibility and state the requirements for each. We also organize open-weight reasoning models by model-side and interface mechanisms. Our empirical study covers broad knowledge, symbolic reasoning, and competition mathematics, and we publicly release 1,403,520 sampled model attempts. The project website is available at https://mohsenhariri.github.io/scorio/tts. The released datasets are Trace (https://huggingface.co/datasets/harimo/scorio-trace), Lite (https://huggingface.co/datasets/harimo/scorio-lite), Math (https://huggingface.co/buckets/harimo/scorio-math), and SuperGPQA (https://huggingface.co/buckets/harimo/scorio-gpqa).

On-policy training has emerged as a powerful post-training paradigm for improving the reasoning capabilities of large language models, and is often enhanced by golden trajectories from stronger expert models. However, when the expert fails on harder problems, existing trajectory-guided methods lose their main source of supervision, and these failed trajectories are typically discarded as negative samples. We argue that such failures, which we call Golden Negative Trajectories, can still provide valuable reasoning signals when treated not as demonstrations to imitate, but as flawed trajectories to reflect upon. We identify a Reflection Advantage: for hard problems, reflecting on a flawed trajectory can be easier and more effective than solving the problem directly from scratch. Motivated by this, we propose ReflectRL, a lightweight plug-and-play framework that learns from Golden Negative Trajectories during on-policy training. ReflectRL first uses these trajectories to elicit Reflective Reasoning, then applies Reflective-to-Direct Policy Transition to transfer the acquired reasoning behavior back to Direct Reasoning. Experiments across 9 benchmarks, 4 LLM backbones, and 4 on-policy training methods show that ReflectRL consistently improves reasoning performance with minimal overhead.

When a language model fails on surface-perturbed input (typos, OCR noise, homophones), "which layer is responsible" has three natural operationalizations: where representations diverge most (sensitivity), where restoring clean activations recovers the prediction (causality), and where a small adapter can repair the damage (compensatory capacity) - and we show these three layer maps dissociate. Across a five-model panel we identify two propagation regimes - spike-and-suppress (Phi-3.5, Gemma-2-9B) and late-accumulation (Llama-3, Mistral, Qwen2.5-7B) - and on the two models meeting an 80% identity-patch gate, sensitivity and causality are anti-correlated (rho = -0.72 to -0.88). Within-family scaling on Qwen2.5 (1.5B to 14B) shows the late-accumulation signature strengthening monotonically with scale, corroborated on a second family. We propose cascade disruption as the mechanism behind the dissociation: adapters placed at causally implicated early layers break intact downstream computation, making diagnostic-flagged sites the worst adapter placements. A fixed-harness layer sweep across four models (3.8-8B) confirms the core prediction on chain-of-thought GSM8K - the flagged sites are the most damaging adapter windows on every adjudicable model - and is sign-consistent but strongly attenuated on a multiple-choice control, consistent with damage that compounds with generation length. The sweep yields practical guidance: a training-free LRD pre-screen and a default-deepest placement rule, though absolute gains over no-adapter baselines remain small. Finally, apparent gains from a representation-stability loss reverse under an adequate generation budget - truncated chain-of-thought had been scored as empty - a methodological warning for any intervention evaluated on chain-of-thought tasks.

Chain-of-Thought (CoT) reasoning offers a promising window into model monitoring. However, monitoring relies on faithfulness, i.e., the model output strictly derives from its reasoning trace. We identify an alignment tension where a model must be faithful enough to be monitored, yet robust enough to reject unsafe reasoning. We demonstrate that this counterbalance exists in current Large Reasoning Models (LRMs), and show ways in which it can be addressed. We introduce HazMart, a human-written dataset set in an autonomous AI shopkeeper scenario. Unlike prior work that relies on providing hints in prompts to test faithfulness (e.g., "A Stanford professor said it should be Answer A"), we propose a novel replacement-based technique, which we call Targeted Reasoning Replacement (TRR), that directly intervenes in the reasoning chain to substitute in unsafe or illogical thoughts (e.g., "Wait, the answer must be Option B [was Option A] because it is the most fitting"). DeepSeek-R1-Llama-70B exhibits high faithfulness (97.5%) but fails to reject Unsafe Reasoning (12.3%), while QwQ-32B is more robust (73.9% safety) at the cost of lower faithfulness (74.7%). Mechanistic analyses of QwQ-32B reveal that these properties are represented by anti-correlated internal directions peaking at the action-commit token. Finally, we demonstrate that representation steering can independently amplify the safety direction, increasing safe behavior by 9 percentage points while maintaining base capabilities.

Chain-of-Thought (CoT) prompting remains the standard baseline for evaluating models' reasoning abilities. Originally, this technique was introduced to elicit step-by-step reasoning from large language models (LLMs), which would otherwise tend to directly output the final answer. However, many modern LLMs produce CoT-style responses \textit{natively} when presented with reasoning tasks, which made us revisit the effectiveness of standard CoT prompting. We evaluate several modern mid-sized language models on a math problem-solving task and find that models specialized for reasoning achieve better performance in a simple zero-shot setting than when using few-shot CoT examples - significantly surpassing officially reported results at no additional cost (e.g., from $\sim$77% to $\sim$84% for Mathstral on GSM8K). For the tested general-purpose model, a zero-shot CoT prompt is also sufficient to outperform a few-shot CoT baseline. We attribute this to a `guidance-distraction' tradeoff: standard CoT prompting also demands style adaptation, formatting compliance, and potentially undesired contextualization, which can distract models from the core reasoning task. Our findings suggest that using standard CoT prompting increasingly acts as a source of distraction as models grow stronger.

Self-consistency assumes the most frequent answer among sampled reasoning traces is the most reliable, but this can fail in causal reasoning: samples often repeat the same confounding error, and votes fragment across multiple valid answers, letting an invalid answer win despite a valid minority trace. We introduce CALVER (Causal Axiom-Level VERification), a training-free symbolic verifier that scores structured traces against Pearl's causal criteria, including -separation, backdoor adjustment, and intervention, and selects the highest-scoring candidate without consulting a reference answer. On CLEAR find-one-valid queries that admit multiple graph-valid answers, CALVER reaches 42.1% where plurality, a reward model, an LLM judge, and model confidence remain near 30% on identical frozen pools. Scaling the judge to 72B does not close the gap. In an audited clean-core subset, 11 of 21 graph-valid CALVER selections differ from the benchmark's listed answer while still satisfying the requested predicate. The advantage widens with the sampling budget and reproduces across ten published Bayesian networks, a second model family, and settings where the model must build the graph from text. CALVER also improves thresholded average-treatment-effect decisions against exact ground truth, generalizes to logic under a truth-table checker, and scores each candidate in milliseconds on CPU. CALVER needs only a causal structure, supplied outright or built from the text; wherever that holds, selection can aggregate via causal validity.

Reasoning in Multimodal Large Language Models (MLLMs) requires both fine-grained visual perception and rigorous logical deduction. Explicit text-based Chain-of-Thought (CoT) is computationally expensive and prone to visual hallucinations, while existing latent reasoning methods typically require costly training. Furthermore, directly adapting training-free LLM reasoning mechanisms to the multimodal setting yields unstable performance. We identify that this failure stems from their reliance on token-level entropy, which fundamentally conflates perceptual ambiguity (e.g., unclear visual details) with logical uncertainty (e.g., complex reasoning steps). To overcome this bottleneck, we present a novel training-free inference strategy for MLLMs that explicitly decouples perception and reasoning. We propose a novel metric, the vision-to-text attention ratio, to dynamically gauge the model's cognitive focus. Guided by this metric, our proposed framework, Attention-Guided Switching (AGS), adaptively triggers latent reasoning for perceptual tokens to preserve high-fidelity visual information in the continuous space, while enforcing explicit text generation for logical tokens to maintain structural anchoring. Extensive experiments demonstrate that our method achieves state-of-the-art performance, significantly improving both accuracy and inference efficiency by reducing autoregressive steps and latency. Code is released at https://github.com/swordAndSnow/MM26-AGS.

Large language models often reason at length before answering, increasing cost and latency. Prompts and trained settings can shorten this reasoning, but a shorter trace may only show that the model stopped sooner. Here, we evaluate paired runs of the same question at matched reasoning horizons across 198 GPQA Diamond and 500 MMLU-Pro questions. We test a numeric/concision prompt that announces a token limit for Qwen3-14B and the trained effort settings of gpt-oss-20b and -120b. The Qwen prompt shortens reasoning traces by 12-17%, while accuracy changes at matched token limits are small and mixed. A concise/early-answer instruction raises MMLU-Pro accuracy by 3.8 percentage points at 512 tokens, including +2.7 points when both runs are unfinished. Its gain at 2,048 tokens is uncertain. For gpt-oss, candidate-logit answers from completed low- and medium-effort reasoning are 14.5-26.3 points more accurate than matched-horizon high-effort answers. Most of the 512-token advantage comes from lower effort finishing earlier, while differences among unfinished runs are smaller and mixed. Wrong early answers often concentrate probability on the chosen option, so earlier stopping does not uniformly improve probability quality. In these tests, a tight deadline can favor lower effort or a concise instruction, whereas allowing high effort to finish can recover higher final accuracy. Evaluations should report correct completion before a deadline, the answer obtained when a run is stopped, differences among unfinished runs, and probability assigned to the correct answer separately.

Long Document Visual Question Answering (LongDocVQA) requires Multimodal Large Language Models (MLLMs) to locate, integrate, and reason over heterogeneous document elements distributed across multiple pages. Existing approaches, including end-to-end MLLMs, retrieval-augmented generation (RAG) pipelines, and document agents, often lack explicit mechanisms to represent and verify how grounded evidence is progressively composed during reasoning, limiting both answer accuracy and traceability. In this paper, we cast LongDocVQA as an explicit evidence graph reasoning problem rather than implicit answer prediction. To this end, we propose DocTrace, a hierarchical framework that progressively performs evidence localization, structured document parsing, and evidence graph reasoning to enable explicit evidence provenance. To effectively learn these capabilities, we develop a two-stage training framework: joint Supervised Fine-Tuning (SFT) first initializes evidence localization and graph reasoning abilities, followed by task-specific Group Relative Policy Optimization (GRPO) with dedicated rewards to further optimize these capabilities. Extensive experiments on MMLongBench-Doc, LongDocURL, and SlideVQA demonstrate that DocTrace consistently outperforms both existing open-source baselines and proprietary MLLMs. Compared with the Qwen3-VL-8B-Instruct backbone, DocTrace achieves absolute improvements of 14.4, 11.3, and 11.7 points on the three benchmarks, respectively. Beyond competitive performance, DocTrace constructs traceable evidence graphs with explicit node-level provenance, enabling transparent and verifiable reasoning for long document understanding.

Chain-of-thought (CoT) reasoning improves large language model (LLM) performance while also providing an observable interface to the model's reasoning process. Existing approaches that leverage verbalized CoTs to monitor reasoning correctness, however, largely evaluate the semantic correctness or consistency of individual intermediate steps, rather than how the reasoning process evolves across the trace. As a result, failures distributed across the reasoning trajectory, rather than those localized to a single incorrect step, remain comparatively underexplored. Furthermore, verbalized CoTs need not faithfully reflect the model's internal reasoning, motivating analyses that do not treat individual statements as literal accounts of internal computation. In this work, we therefore ask whether the dynamics of visible CoT can be leveraged to systematically distinguish successful from failed reasoning without assuming such semantic faithfulness. We study a range of LLMs on verifiable Boolean satisfiability tasks with variable complexity, enabling controlled comparisons near each model's capability frontier. Tagging CoT sentences by reasoning function reveals premature verification collapse on SAT problems: incorrect traces enter clause checking earlier, repeat similar operations, and finalize sooner. On UNSAT problems, models presumptuously move towards incorrect SAT conclusions, checking candidate assignments rather than deriving contradictions across constructed cases. Subsequently, a targeted proof-search prompt intervention raises Llama3-70B accuracy from 13.3% to 85%, correcting 84.6% of these errors. These results show that capability failures can manifest as distributed, task-dependent changes in the structure of visible reasoning, and that CoT dynamics agnostic to whether the verbalized trace reflects the model's internal computations can help diagnose and correct failures.

Parameter-efficient fine-tuning (PEFT) enables efficient adaptation of large language models, but existing MoE-based PEFT methods typically improve capacity by storing multiple full LoRA experts, causing adapter storage to grow linearly with the number of experts and restricting adaptation to a fixed expert pool. We ask whether MoE-based PEFT can produce instance-specific adaptations without explicitly storing a separate LoRA module for each expert. To address this gap, we propose MoEGen, an adaptation framework that shifts MoE-based PEFT from expert selection to expert-conditioned parameter generation. Instead of storing each expert as a full LoRA adapter, MoEGen represents each expert as a small learnable vector, termed an expert code. It routes each input over these vectors and uses their weighted combination to condition a lightweight hypernetwork that generates input-specific low-rank updates. This design decouples expert capacity from adapter storage while enabling instance-conditioned adaptation. Experiments on eight commonsense reasoning benchmarks show consistent improvements over strong static and MoE-based PEFT baselines across three backbones. MoEGen also performs strongly in joint medical and legal-domain adaptation.

Reinforcement Learning with Verifiable Rewards (RLVR) improves LLM reasoning but typically relies on ground-truth (GT) answers, limiting scalability. Voting-based label-free RLVR replace gold supervision with answer-level consensus from model samples. However, collapse arises when the same answer-level signal is used both to estimate rewards and to drive token-level policy optimization, encouraging the model to directly reinforce answer tokens rather than improve reasoning. We propose OM-GRPO, a label-free RLVR framework that decouples reward estimation from policy optimization. OM-GRPO masks gradients on the answer span while retaining answer-level rewards through a soft consensus signal, shifting optimization pressure away from answer tokens. We further introduce Contrast-Augmented Reward, which refines reward estimation via low-cost pairwise comparisons over existing trajectories without additional rollouts. Across diverse reasoning benchmarks and three LLM backbones, OM-GRPO consistently outperforms existing label-free RLVR methods and matches supervised GT-reward training with stable optimization. This stability is particularly beneficial in the Test-Time Training setting, where OM-GRPO surpasses majority voting by 4.24 points.

Multi-domain machine translation (MDMT) requires more than fluent generation: it demands domain-sensitive translation decisions such as domain disambiguation, terminology control, and stylistic adaptation. Large reasoning models (LRMs) make such decisions explicit through intermediate translation steps, but our analysis across 15 domains and four translation directions shows that this explicit reasoning is double-edged: it improves long-form and high-difficulty translation, yet often drifts in terminology-intensive and stylistically constrained settings. We trace this failure to a credit-assignment bottleneck: existing methods optimize final outputs or coarse trajectories, but cannot identify which translation steps actually help the final translation. To address this, we propose PAMT, a process-aligned training framework that combines cold-start domain-aware Long-CoT supervision with reinforcement learning. PAMT uses sequence-level format and outcome rewards for the final translation, together with a step-level process reward that measures how much each explicit translation step increases the likelihood of the reference translation. Across two backbones, PAMT improves over base models, outperforms MT-specialized baselines on average, and remains competitive with strong LLMs/LRMs across in-domain, OOD, and multilingual settings.

Reinforcement learning (RL) has emerged as an effective method for enhancing the reasoning capabilities of large language models (LLMs). However, existing methods suffer from insufficient precision in feedback on generated answer trajectories and exhibit the phenomenon of problem difficulty drift. To address these challenges, we propose CVPO - Curriculum-guided Value-Variance Policy Optimization. At the response trajectory level, we find that token-level value-variance correlates with exploration intensity. Our theoretical analysis shows this variance bounds policy update magnitude. We then use the estimated trajectory value-variance to quantify the intrinsic randomness in generation. Based on this, we design a variance-aware advantage adjustment mechanism for different reward types. At the question level, we introduce a dynamic curriculum weighting method that adapts to question difficulty. This helps the model focus on tasks matched to its current ability during each training stage. Experimental results show our method outperforms strong value-based baselines like VAPO. It achieves better performance and stronger exploration, enabling more accurate and robust reasoning in language models across various math tasks.

Long-context reasoning remains a critical bottleneck for large language models, as recent recurrent-memory approaches face two inherent challenges: sequential chunk-wise updates can overwrite early critical evidence with later irrelevant content, and serial inter-chunk dependencies limit parallelism and cause latency to increase with context length. To address these issues, we propose PI-Mem (Parallel-Iterative Memory), a mechanism that processes all chunks in parallel and iteratively refines a shared memory over a bounded number of turns. In each turn, PI-Mem reads all chunks in parallel conditioned on the current memory, selects new or complementary evidence from each chunk, and merges the selected evidence into a compact shared memory for the next turn. To discourage redundant turns, we optimize the workflow through reinforcement learning with an auxiliary turn-efficiency reward, enabling the model to adaptively exit once sufficient evidence has been accumulated. We evaluate PI-Mem with Qwen3.5-35B-A3B and Qwen2.5-7B on the HotpotQA benchmark across context lengths up to 3.6 million tokens and find that it outperforms the recurrent-memory baseline by +6.25 and +7.81 absolute points while achieving 6.1$\times$ and 2.1$\times$ inference speedups, respectively. These results demonstrate that PI-Mem breaks the accuracy--efficiency trade-off in long-context reasoning and provides a scalable approach to complex multi-hop question answering over extremely long documents.

Large language models (LLMs) have demonstrated impressive performance in MQM-based translation quality (TQ) evaluation, and recent advances in large reasoning models (LRMs) promise even greater improvements. However, both LLMs and LRMs are computationally expensive to deploy at scale, while small language models (SLMs)---though much more efficient---struggle with the complex reasoning required for evaluation tasks. In this work, we present an extensive empirical study benchmarking SLMs, LLMs, and LRMs across a wide range of TQ evaluation setups, providing a comprehensive view of the current landscape and establishing best practices. To address the scalability challenge, we introduce TQLite, a novel distillation framework that enables SLMs to approach the MQM evaluation performance of the best LRM-based evaluators. Our approach leverages a multi-LRM jury to generate high-quality synthetic training data via practical data curation techniques and aggregation of evaluation responses across a diverse panel of models. Our results demonstrate that SLMs trained via TQLite achieve strong MQM evaluation performance that far exceeds off-the-shelf evaluation capabilities of standard SLMs, offering a scalable and cost-effective alternative to LLM- and LRM-based evaluators.

Although reinforcement learning with verifiable rewards (RLVR) has improved the performance of large language models (LLMs) across a variety of reasoning tasks, there is significant debate as to whether RLVR expands the reasoning capability boundary, or just improves sampling efficiency. In this paper, we investigate the nature of test-time exploration in RLVR-trained LLMs by employing controlled maze-solving experiments and extracting a tree structure from mathematical reasoning traces (BODHI-Trees) based on semantic equivalence. This helps us delineate between entropy arising from stylistic variations and genuine inferential branching. Our findings demonstrate that the policy entropy collapse observed in RLVR models is not merely syntactic, and is accompanied by a significant reduction in semantic branching entropy. While RLVR improves adherence to environmental constraints and backtracking capabilities, it constricts the space of continuations; we provide evidence suggesting that this might be responsible for the sample efficiency gains of RLVR, albeit at the cost of genuine rollout diversity.

Chain-of-thought (CoT) monitoring is an increasingly important component of AI safety stacks but relies on the assumption that a model's reasoning trace is informative about its actions. This work studies the limits of CoT monitoring through the lens of model poisoning. We demonstrate that backdoors can be implanted into reasoning models to elicit an attacker-chosen behavior while their CoT traces appear entirely benign. We find that these CoT-Hidden backdoors can be induced through simple fine-tuning recipes across reasoning-model architectures and sizes. When direct poisoning is ineffective, we introduce a curriculum training approach that progressively teaches the model to produce an attacker-chosen output while concealing the behavior from its reasoning traces. These findings suggest that CoT monitoring may be better framed as a question about the consistency between a model's reasoning trace and its final response than as anomaly detection within a trace. We further examine the mechanisms that allow models to suppress evidence of the target behavior from their reasoning traces. Causal interventions locate a trigger-conditioned activation pathway that does not depend on the visible reasoning, and residual stream verbalizations provide an anomaly warning near answer generation, but do not identify the trigger, target, or backdoor mechanism.

Scientific reasoning benchmarks typically evaluate large language models (LLMs) using final-answer accuracy. However, a correct answer does not necessarily demonstrate the reasoning capability targeted by the problem. We identify Solution Hacking, a failure mode in which an LLM reaches the correct answer through invalid shortcuts, such as numerical search, enumeration, guessing, or answer-first verification, without providing a valid task-targeted derivation. We systematically analyze this phenomenon across difficulty levels, scientific domains, and frontier models. Solution hacking increases sharply with benchmark difficulty, from 2.2% on common problems to 28.3% on Olympiad-level problems and 37.4% on HLE. Moreover, 8.2%-44.1% of answers credited as correct across frontier models are identified as hacked solutions. We further develop expert-inspired anti-hacking strategies, including an automatic judge and a test-time instruction. The results show that suppressing shortcut behavior substantially reduces reported accuracy while having a smaller effect on correct and non-hacked accuracy. These findings reveal that answer-only evaluation can overestimate the scientific reasoning capabilities of frontier LLMs.

Reinforcement learning has become a central paradigm for improving the reasoning capabilities of large language models. Existing methods generally aim to reduce the failure probabilities induced across problems. In this paper, we introduce a moment-based perspective on policy optimization for LLM reasoning by treating the failure probability of a randomly sampled problem as a random variable and characterizing optimization objectives through its moments. Under this perspective, many existing methods optimize only a single moment of the failure-probability distribution, leaving its broader distributional structure largely uncharacterized. We propose \textbf{M}ulti-\textbf{M}oment \textbf{P}olicy \textbf{O}ptimization (MMPO), a novel policy optimization framework that jointly minimizes multiple moments of the failure-probability distribution. MMPO admits a direct operational interpretation as minimizing the expected truncated time required to obtain the first successful response. Beyond MMPO, we further develop a general moment-transformation framework that systematically induces different moment profiles and provides a unified view of a broader family of policy optimization objectives. Experiments across five mathematical reasoning benchmarks and models of different scales demonstrate that MMPO consistently outperforms strong baselines. We hope this moment-based perspective offers new insights into the design of policy optimization objectives for LLM reasoning.

Large language models often show users a final response and a short reasoning summary while the full reasoning trace stays hidden. We introduce an observability ladder that holds each completed run fixed and varies only what a reader inspects to judge whether the answer is correct: the response, a self-summary the model writes from the trace, the trace itself, and internal signals, each with and without the prompt. Across three benchmarks and five open-weight Qwen3 and gpt-oss models, we train matched linear correctness predictors on each access level. Without the prompt, summaries carry most of the trace's ranking signal (mean AUROC 0.774 versus 0.813) and add +0.156 over the response alone. With the prompt visible, the summary's gain collapses to +0.019, while the trace still adds +0.041. Even at equal length, the trace's last words predict correctness as well as summaries, or slightly better, and carry denser and more discriminative uncertainty and self-correction cues. On MMLU-Pro questions with both correct and incorrect runs, linear summary readers are near chance and trace readers retain only modest signal, both with and without the prompt (prompt-withheld AUROC 0.503-0.545 versus 0.544-0.590). With the prompt withheld, a GPT-5-mini reader recovers substantially more signal from both summaries and traces on gpt-oss-20b, and even then the trace keeps a small +0.034 advantage. Much of the linear readers' trace signal is associated with length. In the common case where users already hold the prompt, summaries are less helpful than the full trace for monitoring correctness. Monitorability is thus a joint property of the display and the reader, so any monitorability claim, including for faithfulness, should specify both.

Reflection is a powerful mechanism for LLM reasoning, yet its effectiveness hinges on accurately attributing failures to specific reasoning steps, a capability that current models notably lack. Existing failure attribution methods either require expensive step-by-step counterfactual testing that scales poorly with trajectory length, or treat reasoning traces as flat sequences that ignore the inherent non-linear logical dependencies. We propose a hypergraph-based paired failure attribution (HPFA) framework that attributes the failure root cause by comparing the hyperedges of the targeted failure reasoning path against a reference successful path. By reducing the search space, our method efficiently localizes root causes and enables scalable synthesis of attribution data for training a lightweight attributor model via supervised fine-tuning and reinforcement learning. Experiments on mathematical reasoning and agentic coding tasks demonstrate that HPFA can dramatically increase attribution accuracy and efficiency, and the trained attributor consistently improves reasoning accuracy at test time, outperforming baselines that lack graph structure or paired analysis.

Chain-of-thought (CoT) reasoning can improve performance on difficult video questions but often wastes decoding tokens on simple ones. We study whether a video multimodal large language model can adapt its reasoning effort to each question. We propose AdaThinkV, an adaptive framework for video reasoning that learns whether to reason explicitly without offline difficulty labels, manually tuned confidence thresholds, or an external router. During reinforcement learning, AdaThinkV samples matched rollouts in explicit reasoning and direct answering modes for each prompt. ThinkGain estimates the prompt-level utility of explicit reasoning by balancing its accuracy gain against additional response length, providing supervision for both conditional response generation and autonomous mode selection. For difficult prompts, limited rollout exploration can yield groups in which every response is unsuccessful and accuracy rewards show little variation, providing insufficient signal for learning. We therefore introduce Variance Recovery Policy Optimization (VRPO), which retains and progressively expands these groups to recover informative signals from prompts that are difficult yet solvable. At inference, AdaThinkV selects a response mode and generates the response in a single autoregressive sequence. Across a unified suite of video reasoning evaluations, AdaThinkV achieves a mean accuracy of 40.79 with an average of 257.20 output tokens, outperforming the strongest evaluated adaptive baseline by 2.98 points while using 22.7% fewer tokens. Project page: https://trilarflagz.github.io/AdaThinkV/

Multimodal Large Reasoning Models (MLRMs) have achieved strong performance on tasks requiring visual understanding and multi-step inference. However, as reasoning trajectories grow, models may become less effective at using information established earlier in the context, increasing the risk of reasoning errors. Existing approaches primarily address this problem by sustaining visual grounding throughout reasoning. However, reasoning also transforms visual observations into task-specific relations, constraints, and intermediate conclusions whose influence may weaken over long trajectories. Our attribution analysis suggests that correctness is not consistently separated by image attribution alone, but is more closely associated with whether trajectories retain and integrate such reasoning-derived information across stages. Motivated by this, we introduce TRAM (TRajectory-derived Auxiliary Memory), a training-free method that augments standard decoding with an auxiliary memory pathway derived from the model's own reasoning trajectory. TRAM consolidates completed reasoning into a compact latent memory, updates it online through fast and slow recurrent streams, and feeds it back into selected decoder layers through a lightweight residual pathway. Experiments across four MLRM variants on eight benchmarks show that TRAM improves performance over vanilla decoding on mathematical, scientific, and general visual reasoning tasks without additional training.

Most time series (TS) models are specialized for a single task, either understanding (i.e., returning text answers about a TS) or generation (i.e., returning a numeric forecast). Only recently have unified models begun to handle the two within a single architecture. Even these models, however, produce the two outputs as task-separated paths and cannot predict a series and explain why that prediction arises within a single coherent response. In this paper, we argue for a task-fused model that jointly produces 1) prediction (generation) and 2) selfexplanation (understanding), thereby integrating 1) numerical TS forecasting and 2) interpretable text reasoning within a single response. To enable the systematic study of this capability, we present both a benchmark and a recipe that jointly address the two tasks. The benchmark, ReasonTS-Bench, identifies five fundamental patterns underlying TS and enables the joint evaluation of both tasks. ReasonCast, our recipe for finetuning any LLM to perform both tasks jointly, yields a model that generates a reasoning chain and a forecast together in a single autoregressive pass. Extensive experiments show that ReasonCast outperforms both LLMs and TS models on prediction accuracy while producing verifiable, causal reasoning. Code is available at: https://github.com/seunghan96/reasoncast.

Multimodal large language models (MLLMs) have advanced visual understanding and reasoning, yet their static parametric knowledge limits their ability to address knowledge-intensive and dynamically evolving open-world problems. To move beyond this limitation, multimodal deep search has emerged as a key direction for open-world information access, evolving from single-turn factual retrieval toward long-horizon, multi-turn search guided by visual evidence. However, existing methods typically confine vision to the input or answer stage, overlooking its role in intermediate reasoning, and lack designs tailored to long-horizon interaction. Consequently, visual evidence rarely drives continued retrieval, constraining both interaction depth and reasoning span. To address these limitations, we propose DeepVoyager-VL, a long-horizon multimodal deep-search framework for vision-in-the-loop search. Specifically, we construct a multimodal event graph to drive data synthesis, yielding problems with intermediate visual dependencies and long reasoning chains. We then design an agent framework for active visual acquisition and on-demand image loading. Finally, we fine-tune models on the synthesized data without reinforcement learning. Extensive experiments across ten multimodal search benchmarks demonstrate the effectiveness of our method.

Verifier-guided reinforcement learning has become a powerful paradigm for improving LLM reasoning. In multi-turn settings, models receive a verifier score after each turn and iteratively refine their outputs. Although such scores provide dense feedback, they do not directly provide dense credit: a score measures the quality of the current output, while credit should measure how the current turn changes the refinement trajectory. We propose TCPO, a turn-level credit assignment method for verifier-guided multi-turn RL. TCPO casts credit assignment as score-to-credit conversion and constructs turn-level advantages through reference-based comparisons: retrospective credit captures immediate progress and regression relative to the best prior state; hindsight delayed credit identifies non-improving turns with later payoff; and selective fixed-history counterfactual estimation refines high-surprisal turns under the same history. Experiments on math reasoning, code generation, and AppWorld agent tasks show that TCPO improves or matches the strongest baselines across model scales, task domains, and verifier types. TCPO achieves the best or tied-best best-turn Pass@8 on Qwen3-4B and DeepSeek-R1-Distill-Llama-8B, reduces turns to success, and improves multi-turn agent performance. These results highlight score-to-credit conversion as a central ingredient for verifier-guided multi-turn policy optimization.

KV cache compression is commonly evaluated by final-answer accuracy, implicitly assuming that preserving the answer also preserves the reasoning that supports it. We test this assumption for large reasoning models and show that it can fail: under compression, correct answers and the validity of their visible supporting rationales can be preserved at different rates. We study this failure with a controlled fixed-trace replay protocol, which holds reasoning content fixed and isolates whether compression preserves usable information from an already available trace. We evaluate ten token-eviction KV compression methods and one quantization method on three models across mathematical reasoning, scientific QA, clinical calculation, and long-context retrieval. We measure final accuracy, answer-chain consistency, and perturbation faithfulness. Across tasks, token-eviction methods can preserve competitive final-answer accuracy while substantially degrading chain support or perturbation faithfulness. We call this the answer-evidence gap. A coverage-preserving quantization control is substantially less affected, suggesting that the failure is tied less to KV memory reduction itself than to losing access to parts of the reasoning trace. Code is available at https://github.com/famous-blue-raincoat/Safe_KV_Compress.

Simulating human-like Theory of Mind (ToM) has been a longstanding problem in natural language processing (NLP). To address this, existing works introduce a reasoning step of event hiding (a.k.a. perspective-taking), where events unknown to a character are removed before question answering. However, resorting to event hiding for ToM reasoning presents a performance degradation issue due to the strict output format constraints involved in event hiding. To mitigate this issue, we propose generating perspective-taking outputs as free-form explanations without event hiding, but this poses a notable yet underexplored challenge: LLMs need to inhibit responses to events unknown to characters, because the absence of event hiding exposes LLMs to these events throughout reasoning. To address this challenge, we hypothesize and empirically verify that LLMs can achieve such inhibition if a character's lack of knowledge about events is made explicit during reasoning. Based on this finding, we introduce PICTURE, a new prompting method that enables LLMs to generate a character's lack of knowledge within free-form Chain-of-Thought (CoT). Experimental results show that PICTURE outperforms existing prompting methods by an average of 7.3% on false-belief tasks.