Language Model Distillation

Follow this topic Ask about this topic

Language model distillation studies how to transfer a teacher model’s knowledge and behavior into a student that is cheaper to train or deploy, while retaining useful capabilities. Common approaches train on teacher-generated answers or reasoning traces, match teacher and student token probabilities, align hidden representations, or compress model weights; newer on-policy methods also use feedback on student-generated text. Researchers evaluate these methods on reasoning, coding, instruction-following, multilingual, and agent tasks, often comparing capability retention against training and inference costs. A central challenge is that teacher guidance can be noisy or poorly matched to a student’s abilities, so improving transfer efficiency must be balanced against stability, generalization, diversity, and loss of existing capabilities.

Latest papers

522 papers135 papers in the last four weeks, up 160% on the four weeks before. 0.8% of all new papers.

Mar 23, 2026

Dual-Space Knowledge Distillation with Key-Query Matching for Large Language Models with Vocabulary Mismatch

Large language models (LLMs) achieve state-of-the-art (SOTA) performance across language tasks, but are costly to deploy due to their size and resource demands. Knowledge Distillation (KD) addresses this by training smaller Student models to mimic larger Teacher models, improving efficiency without significant performance loss. Dual-Space Knowledge Distillation with Cross-Model Attention (DSKD-CMA) has emerged as a SOTA method for KD between LLMs with distinct tokenizers, yet its internal workings remain largely opaque. In this work, we systematically analyse the attention mechanism of DSKD-CMA through manual token alignment probing and heatmap visualisations, revealing both strengths and limitations. Building on this, we introduce a novel method, DSKD-CMA-GA, based on Generative Adversarial (GA) learning, to address the mismatched distributions between the keys and queries computed from distinct models. Experiments show modest but consistent ROUGE-L gains in text generation quality, particularly on out-of-distribution data (+0.37 on average), narrowing the gap between cross- and same-tokenizer KD.

Stella Eva Tsiapali, Cong-Thanh Do, Kate Knill

University of Cambridge, Department of Engineering, Cambridge, United Kingdom · Toshiba Europe Limited, Cambridge, United Kingdom

Cross-Tokenizer Knowledge Distillation Language Model Distillation Adversarial Knowledge Distillation

Mar 22, 2026

Demystifying Low-Rank Knowledge Distillation in Large Language Models: Convergence, Generalization, and Information-Theoretic Guarantees

Knowledge distillation has emerged as a powerful technique for compressing large language models (LLMs) into efficient, deployable architectures while preserving their advanced capabilities. Recent advances in low-rank knowledge distillation, particularly methods like Low-Rank Clone (LRC), have demonstrated remarkable empirical success, achieving comparable performance to full-parameter distillation with significantly reduced training data and computational overhead. However, the theoretical foundations underlying these methods remain poorly understood. In this paper, we establish a rigorous theoretical framework for low-rank knowledge distillation in language models. We prove that under mild assumptions, low-rank projection preserves the optimization dynamics, yielding explicit convergence rates of

O(1/\sqrt{T})

. We derive generalization bounds that characterize the fundamental trade-off between model compression and generalization capability, showing that the generalization error scales with the rank parameter as

O(r(m+n)/\sqrt{n})

. Furthermore, we provide an information-theoretic analysis of the activation cloning mechanism, revealing its role in maximizing the mutual information between the teacher's and student's intermediate representations. Our theoretical results offer principled guidelines for rank selection, mathematically suggesting an optimal rank

r^* = O(\sqrt{n})

where

n

is the sample size. Experimental validation on standard language modeling benchmarks confirms our theoretical predictions, demonstrating that the empirical convergence, rank scaling, and generalization behaviors align closely with our bounds.

Alberlucia Rafael Soarez, Daniel Kim, Mariana Costa +1

Department of Computer Science, University of Brasilia

Language Model Distillation Knowledge Distillation

Mar 20, 2026

Grounded Chess Reasoning in Language Models via Master Distillation

Language models often lack grounded reasoning capabilities in specialized domains where training data is scarce but bespoke systems excel. We introduce a general framework for distilling expert system reasoning into natural language chain-of-thought explanations, enabling compact models to acquire domain expertise and the ability to generate faithful, grounded explanations. Rather than distilling only final outputs, we capture the full reasoning process, transforming opaque expert computations into transparent, step-by-step explanations. We demonstrate this approach in chess, a canonical reasoning domain where language models continue to underperform. Our 4B parameter model, C1, advances from a near-zero baseline to 48.1% accuracy, outperforming all open-source models and most frontier proprietary systems. Notably, C1 surpasses its distillation teacher and generates solutions in two orders of magnitude fewer tokens than baselines. Unlike prior neural chess approaches that predict only best moves, C1 generates explainable solutions revealing strategic reasoning. Our pipeline combines supervised fine-tuning and reinforcement learning with theme-balanced data sampling for comprehensive tactical coverage. Master Distillation demonstrates how to inject expert-level knowledge into compact models for under-optimized domains, offering a recipe for unlocking RLVR where LLMs lack sufficient base capabilities.

Zhenwei Tang, Qianfeng Wen, Seth Grief-Albert +4

Department of Computer Science, University of Toronto · Queen’s University · Coolwei AI Lab

LLM Grounding CoT Distillation Language Model Distillation Chess LLM Reasoning

Mar 19, 2026

Nemotron-Cascade 2: Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distillation

We introduce Nemotron-Cascade 2, an open 30B MoE model with 3B activated parameters that delivers best-in-class reasoning and strong agentic capabilities. Despite its compact size, its mathematical and coding reasoning performance approaches that of frontier open models. It is the second open-weight LLM, after DeepSeekV3.2-Speciale-671B-A37B, to achieve Gold Medal-level performance in the 2025 International Mathematical Olympiad (IMO), the International Olympiad in Informatics (IOI), and the ICPC World Finals, demonstrating remarkably high intelligence density with 20x fewer parameters. In contrast to Nemotron-Cascade 1, the key technical advancements are as follows. After SFT on a meticulously curated dataset, we substantially expand Cascade RL to cover a much broader spectrum of reasoning and agentic domains. Furthermore, we introduce multi-domain on-policy distillation from the strongest intermediate teacher models for each domain throughout the Cascade RL process, allowing us to efficiently recover benchmark regressions and sustain strong performance gains along the way. We release the collection of model checkpoint and training data.

Zhuolin Yang, Zihan Liu, Yang Chen +14

RL for Language Model Reasoning RL for LLM Agents Reinforcement Learning Language Model Distillation On-Policy Distillation

Mar 16, 2026

Effective Distillation to Hybrid xLSTM Architectures

There have been numerous attempts to distill quadratic attention-based large language models (LLMs) into sub-quadratic linearized architectures. However, despite extensive research, such distilled models often fail to match the performance of their teacher LLMs on various downstream tasks. We set out the goal of lossless distillation, which we define in terms of tolerance-corrected Win-and-Tie rates between student and teacher on sets of tasks. To this end, we introduce an effective distillation pipeline for xLSTM-based students. We propose an additional merging stage, where individually linearized experts are combined into a single model. We show the effectiveness of this pipeline by distilling base and instruction-tuned models from the Llama, Qwen, and Olmo families. In many settings, our xLSTM-based students recover most of the teacher's performance, and even exceed it on some downstream tasks. Our contributions are an important step towards more energy-efficient and cost-effective replacements for transformer-based LLMs.

Lukas Hauzenberger, Niklas Schmidinger, Thomas Schmied +7

1NXAI · 2Johannes Kepler University Linz

LLM Compression Language Model Distillation Teacher-Student Learning

Mar 14, 2026

Retrieval-Feedback-Driven Distillation and Preference Alignment for Efficient LLM-based Query Expansion

Large language models have recently enabled a generative paradigm for query expansion, but their high inference cost makes direct deployment difficult in practical retrieval systems. To address this issue, a retrieval-feedback-driven distillation and preference-alignment framework is proposed to transfer retrieval-friendly expansion behavior from a strong teacher model to a compact student model. Rather than relying on few-shot exemplars at inference time, the framework first leverages two complementary types of teacher-generated expansions, produced under zero-shot and few-shot prompting conditions, as supervision signals for distillation and as candidate pools for preference construction. A retrieval-metric-driven strategy is then introduced to automatically form chosen/rejected expansion pairs according to nDCG@10 differences, and Direct Preference Optimization is applied to explicitly align generation preferences with retrieval objectives. Experiments on TREC DL19/20/21 and MIRACL-zh show that the proposed approach preserves strong retrieval effectiveness while substantially reducing inference cost. In particular, the distilled Qwen3-4B model reaches about 97% of the teacher (DeepSeek-685B) model's nDCG@10 performance on DL19, and remains effective on the Chinese MIRACL-zh benchmark, demonstrating strong practicality across both English and Chinese retrieval settings.

Minghan Li, Guodong Zhou

School of Computer Science and Technology, Soochow University, China

Language Model Distillation Efficient Language Model Inference LLM Alignment

Mar 14, 2026

Knowledge Distillation for Large Language Models

We propose a resource-efficient framework for compressing large language models through knowledge distillation, combined with guided chain-of-thought reinforcement learning. Using Qwen 3B as the teacher and Qwen 0.5B as the student, we apply knowledge distillation across English Dolly-15k, Spanish Dolly-15k, and code BugNet and PyTorrent datasets, with hyperparameters tuned in the English setting to optimize student performance. Across tasks, the distilled student retains a substantial portion of the teacher's capability while remaining significantly smaller: 70% to 91% in English, up to 95% in Spanish, and up to 93.5% Rouge-L in code. For coding tasks, integrating chain-of-thought prompting with Group Relative Policy Optimization using CoT-annotated Codeforces data improves reasoning coherence and solution correctness compared to knowledge distillation alone. Post-training 4-bit weight quantization further reduces memory footprint and inference latency. These results show that knowledge distillation combined with chain-of-thought guided reinforcement learning can produce compact, efficient models suitable for deployment in resource-constrained settings.

Alejandro Paredes La Torre, Barbara Flores, Diego Rodriguez

Duke University

Language Model Distillation LLM Compression

Mar 11, 2026

PACED: Distillation and On-Policy Self-Distillation at the Frontier of Student Competence

Standard LLM distillation treats all training problems equally -- wasting compute on problems the student has already mastered or cannot yet solve. We empirically show that this inefficiency has a precise gradient-level signature: the cross-problem gradient signal-to-noise ratio (SNR) follows a bell curve over student pass rate, collapsing at both extremes. We propose PACED, which weights each problem by

w(p) = p(1{-}p)

where

p

is the student's empirical pass rate -- concentrating training on the zone of proximal development. This requires only student rollouts, no architectural changes, and no hyperparameters. We prove the Beta kernel

w(p) = p^α(1{-}p)^β

is the leading-order optimal weight family arising from the SNR boundary-collapse structure, and is minimax-robust under misspecification (worst-case efficiency loss

O(δ^2)

). Across Qwen3, Qwen2.5, and Llama-3 families, PACED sets a new state of the art in our experimental setting on MATH-500, AIME~~2024, and AIME~~2025, improving over unweighted distillation by up to

\mathbf{+8.2}

and over the strong AKL baseline by up to

\mathbf{+3.6}

, while reducing forgetting to

\mathbf{1.4\%}

and

\mathbf{0.6\%}

in distillation and self-distillation. A two-stage forward-then-reverse KL schedule pushes gains further to

\mathbf{+5.8}

over standard forward KL on the hardest benchmark.

Yuanda Xu, Hejian Sang, Zhengze Zhou +2

LinkedIn Corporation

Language Model Distillation Adaptive Knowledge Distillation Knowledge Distillation On-Policy Self-Distillation

Mar 11, 2026

From Images to Words: Efficient Cross-Modal Knowledge Distillation to Language Models from Black-box Teachers

Knowledge distillation (KD) methods are pivotal in compressing large pre-trained language models into smaller models, ensuring computational efficiency without significantly dropping performance. Traditional KD techniques assume homogeneity in modalities between the teacher (source) and the student (target) models. On the other hand, existing multimodal knowledge distillation methods require modality-specific pre-training of the teacher model, which is computationally infeasible in most cases. In this paper, we introduce ARMADA, an efficient cross-modal knowledge distillation framework designed to transfer knowledge from large vision-language models, including black-box models, to language-only models. Unlike existing KD techniques that rely on the internal structures of multimodal teachers or require computationally expensive pre-training, ARMADA leverages novel alignment techniques to distil knowledge without altering the teacher model, ensuring efficiency and scalability. We empirically validate ARMADA on twelve natural language understanding, eight complex generative reasoning and five instruction-tuning tasks, demonstrating consistent performance improvements in large models such as DeBERTa-v2-1.4B, OPT-1.3B, LLaMA-{3B, 7B, 8B}. ARMADA achieves up to 3.4% improvement on language understanding tasks and 2.6% boost in generative reasoning, all without requiring expensive multimodal pre-training or fine-tuning of the teacher model. Our findings challenge conventional knowledge distillation paradigms by demonstrating that even vision-language models, despite lacking direct textual understanding, can significantly enhance language models when distilled appropriately.

Ayan Sengupta, Shantanu Dixit, Md Shad Akhtar +1

Indian Institute of Technology Delhi, India · Indraprastha Institute of Information Technology Delhi, India

Cross-Modal Knowledge Distillation Knowledge Distillation Language Model Distillation VLM Distillation

Mar 11, 2026

HEAL: Hindsight Entropy-Assisted Learning for Reasoning Distillation

Distilling reasoning capabilities from Large Reasoning Models (LRMs) into smaller models is typically constrained by the limitations of rejection sampling. Standard methods treat the teacher as a static filter, discarding complex "corner-case" problems where the teacher fails to explore valid solutions independently, thereby creating an artificial "Teacher Ceiling" for the student. In this work, we propose Hindsight Entropy-Assisted Learning (HEAL), an RL-free framework designed to bridge this reasoning gap. Drawing on the educational theory of the Zone of Proximal Development (ZPD), HEAL synergizes three core modules: (1) Guided Entropy-Assisted Repair (GEAR), an active intervention mechanism that detects critical reasoning breakpoints via entropy dynamics and injects targeted hindsight hints to repair broken trajectories; (2) Perplexity-Uncertainty Ratio Estimator (PURE), a ratio-based filtering heuristic that reduces high-anomaly shortcut-like rationales; and (3) Progressive Answer-guided Curriculum Evolution (PACE), a three-stage distillation strategy that organizes training from foundational alignment to hard-case adaptation. Extensive experiments on multiple benchmarks demonstrate that HEAL significantly outperforms traditional SFT distillation and other baselines.

Wenjing Zhang, Jiangze Yan, Jieyun Huang +7

Data Science & AI Research Institute, China Unicom · Unicom Data Intelligence, China Unicom

CoT Distillation Language Model Distillation LLM Reasoning Large Reasoning Models

Mar 10, 2026

HoloByte: Continuous Hyperspherical Distillation for Tokenizer-Free Modeling

Sequence modeling universally relies on discrete subword tokenization to circumvent the

\mathcal{O}(N^2)

computational intractability of native byte-level attention. However, this heuristic quantization imposes artificial morphological boundaries, enforces vocabulary dependence, and fractures the continuity of the optimization landscape. To resolve this dichotomy, we introduce HoloByte: a strictly tokenizer-free framework utilizing Continuous Hyperspherical Distillation. HoloByte partitions discrete byte sequences into fixed-capacity chunks and projects them into a continuous, strictly bounded hyperspherical manifold via an invertible, dimension-preserving orthogonal rotation operator. This spatial superposition allows a macroscopic transformer to operate exclusively on compressed continuous representations, formally reducing the exact attention time complexity from

\mathcal{O}(N^2D)

to

\mathcal{O}\left( \frac{N^2}{W^2}D + ND^2 \right)

. A localized causal micro-decoder subsequently unbinds these representations to compute exact byte-level distributions. To govern this continuous trajectory, we propose a dual-objective formulation incorporating a mathematically precise Holographic Latent Mean Squared Error, which strictly bounds the gradient and guarantees asymptotic stability. Theoretically, we derive the minimal embedding dimension

D = Ω(W \ln |\mathcal{V}|)

required to ensure error-free discrete recovery from the continuous manifold. Empirically, under strictly matched parameter constraints, HoloByte is systematically outperforming a comparable discrete Byte-Pair Encoding (BPE) baseline. These results establish Continuous Hyperspherical Distillation as a mathematically rigorous and computationally tractable foundation for vocabulary-invariant sequence modeling. The code is available at https://github.com/VladimerKhasia/HoloByte

Vladimer Khasia

Independent Researcher

Byte-Level Language Model Language Modeling Efficient Attention Language Model Distillation

Mar 9, 2026

LinearARD: Linear-Memory Attention Distillation for RoPE Restoration

The extension of context windows in Large Language Models is typically facilitated by scaling positional encodings followed by Continued Pre-Training (CPT). While effective, this paradigm is notoriously data-hungry and computationally expensive, requiring massive long-text corpora to recalibrate the model to the shifted positional distribution. We propose LinearARD, a restoration distillation method that restores Rotary Position Embedding (RoPE)-scaled students through attention-structure consistency with a frozen native-RoPE teacher. Rather than next-token prediction or opaque hidden-state matching, LinearARD aligns row-wise distributions of dense

Q/Q

,

K/K

, and

V/V

self-relation matrices from the final attention layer to directly supervise attention dynamics. To remove the quadratic memory bottleneck of

n \times n

relation maps, we introduce a linear-activation-memory kernel that stores only per-token log-sum-exp statistics and recomputes logits in the backward pass to obtain exact Kullback-Leibler divergence gradients. Across LLaMA2-7B, LLaMA3-8B, and Mistral-7B-v0.1 extended to 32K context, LinearARD recovers 93.1%/94.2%/94.3% of native short-context performance and achieves strong long-context robustness on RULER using only 4.25M tokens---amounting to just 1.6% (a

\sim

60

\times

reduction) of the 256M-token budget required by state-of-the-art baselines. Under this severely constrained budget, CPT and LongReD remain near zero on RULER, demonstrating that relation-level restoration provides a substantially better efficiency-quality tradeoff. Our code is available at https://github.com/gracefulning/LinearARD.

Ning Yang, Hengyu Zhong, Wentao Wang +4

Institute of Automation, Chinese Academy of Sciences, Beijing, China · Southwest University, Chongqing, China · Dalian University of Technology, Dalian, Liaoning, China +3

Attention-Guided Distillation Memory-Efficient Attention Long-Context Language Modeling Language Model Distillation Knowledge Distillation Efficient Attention+1

Mar 8, 2026

DistillGuard: Evaluating Defenses Against LLM Knowledge Distillation

Knowledge distillation from proprietary LLM APIs poses a growing threat to model providers, yet defenses against this attack remain fragmented and unevaluated. We present DistillGuard, a framework for systematically evaluating output-level defenses against LLM knowledge distillation. We introduce a taxonomy of three defense categories -- output perturbation, data poisoning, and information throttling -- and evaluate nine defense configurations using a standardized pipeline with Qwen3-14B as teacher and Qwen2.5-7B-Instruct as student across three benchmarks (MATH-500, HumanEval+, MT-Bench). Our results reveal that, in a same-family distillation setting against a naive attacker, most output-level defenses are surprisingly ineffective: paraphrasing-based perturbation barely degrades distilled student quality, and data poisoning primarily impairs conversational fluency while leaving task-specific capabilities intact. Only chain-of-thought removal substantially impairs mathematical reasoning (31.4% vs.\ 67.8% baseline), though code generation remains unaffected. These findings demonstrate that the effectiveness of distillation defenses is highly task-dependent and that current output-level approaches are insufficient to broadly prevent knowledge theft.

Bo Jiang

Language Model Distillation Model Extraction Attacks Data Poisoning Attacks

Mar 7, 2026

Unlocking Data Value in Finance: A Study on Distillation and Difficulty-Aware Training

Large Language Models (LLMs) have demonstrated strong general capabilities, yet their deployment in finance remains challenging due to dense domain-specific terminology, stringent numerical reasoning requirements, and low tolerance for factual errors. We conduct a controlled empirical study showing that in specialized vertical domains, performance is largely determined by the quality and difficulty/verifiability profile of post-training data. We introduce ODA-Fin-SFT-318k, constructed via multi-stage distillation and verification to produce high-quality Chain-of-Thought supervision, and ODA-Fin-RL-12k, curated for hard-but-verifiable tasks that balance reward precision and task diversity. Using standard SFT and RL pipelines, we show that high-quality CoT distillation establishes a robust foundation during SFT, while difficulty- and verifiability-aware sampling improves RL generalization. Evaluated on nine benchmarks spanning general financial tasks, sentiment analysis, and numerical reasoning, our ODA-Fin-RL-8B consistently surpasses open-source state-of-the-art (SOTA) financial LLMs of comparable size. We release our ODA-Fin-SFT-318k and ODA-Fin-RL-12k datasets, along with trained models to advance data-centric financial AI research.

Chuxue Cao, Honglin Lin, Zhanping Zhong +5

Shanghai Artificial Intelligence Laboratory, OpenDataLab, OpenDataArena · Hong Kong University of Science and Technology

Financial Language Model Evaluation LLM Training Language Modeling Language Model Distillation Synthetic Data Generation for Language Models RL for Language Models

Mar 7, 2026

Entropy-Aware On-Policy Distillation of Language Models

On-policy distillation is a promising approach for transferring knowledge between language models, where a student learns from dense token-level signals along its own trajectories. This framework typically uses reverse KL divergence, encouraging the student to match the teacher's high-confidence predictions. However, we show that the mode-seeking property of reverse KL reduces generation diversity and yields unstable learning signals when the teacher distribution has high entropy. To address this, we introduce Entropy-Aware On-Policy Distillation. Our key idea is augmenting the standard reverse KL objective with forward KL when teacher entropy is high, capturing the full range of plausible outputs while retaining precise imitation elsewhere. It balances mode-seeking precision with mode-covering robustness without sacrificing on-policy training efficiency. Experiments show that our method maintains generation diversity (sustained token-level entropy) and improves student-teacher alignment (lower forward KL on high-entropy tokens). Across six math reasoning benchmarks, this yields Pass@8 accuracy gains of +1.37 for Qwen3-0.6B-Base, +2.39 for Qwen3-1.7B-Base, and +5.05 for Qwen3-4B-Base compared to baseline on-policy distillation methods. These results demonstrate that accounting for teacher uncertainty is essential for maintaining diversity and achieving effective knowledge transfer.

Woogyeol Jin, Taywon Min, Yongjin Yang +5

KAIST AI, Seoul, South Korea · University of Toronto, Ontario, Canada · Vector Institute, Ontario, Canada +1

Language Model Distillation On-Policy Distillation Uncertainty-Aware Knowledge Distillation LLM Reasoning

Mar 7, 2026

Language-Aware Distillation for Multilingual Instruction-Following Speech LLMs with ASR-Only Supervision

Speech Large Language Models (LLMs) that understand and follow instructions in many languages are useful for real-world interaction, but are difficult to train with supervised fine-tuning, requiring large, task-specific speech corpora. While recent distillation-based approaches train performant English-only Speech LLMs using only annotated ASR data by aligning text and speech using only a lightweight projector, these models under-perform when scaled to multilingual settings due to language interference in the shared projector. We address this by introducing language-aware distillation using a query bank and a gating network that selects or mixes query tokens using a Q-Former projector. Our approach shows gains of 14% over matched multilingual distillation baselines on instruction following. We further synthesize Audio-MLQA, a multilingual spoken QA benchmark built on MLQA with high-quality TTS questions. Our best model improves over existing Speech LLM baselines by 32% on Audio-MLQA.

Shreyas Gopal, Donghang Wu, Ashutosh Anshul +5

College of Computing and Data Science, Nanyang Technological University, Singapore · AI Singapore, National University of Singapore · Institute for Infocomm Research (I2R), A∗STAR, Singapore

Speech Language Models Language Model Distillation Instruction Following

Mar 2, 2026

KDFlow: A User-Friendly and Efficient Knowledge Distillation Framework for Large Language Models

Knowledge distillation (KD) is widely used to compress and post-train large language models (LLMs), yet many existing frameworks execute teacher inference with the same training-oriented backend as student optimization, leading to suboptimal efficiency. In this paper, we propose KDFlow, a novel framework for LLM distillation that features a decoupled architecture and employs SGLang for teacher inference. KDFlow combines SGLang for teacher inference with PyTorch FSDP2 for student optimization, allowing each model to run on a backend tailored to its workload. To enable efficient full-vocabulary distillation in this decoupled architecture, KDFlow transfers the teacher's final hidden states via Ray's object store and recomputes teacher logits on each student worker using a frozen copy of the teacher's output head. Furthermore, our framework supports both off-policy and on-policy distillation and incorporates cross-tokenizer algorithms through highly extensible and user-friendly APIs. Experiments show that KDFlow achieves a 1.44

\times

to 6.36

\times

speedup over MS-SWIFT in off-policy distillation and a 1.43

\times

to 1.75

\times

speedup over verl in on-policy distillation. KDFlow further scales to 64 GPUs, achieving 3.68

\times

and 2.52

\times

strong-scaling speedups in two representative model configurations. The code and documentation are publicly available.

Songming Zhang, Xue Zhang, Tong Zhang +3

School of Computer Science and Technology, Beijing Jiaotong University, Beijing, China · Key Laboratory of Big Data & Artificial Intelligence in Transportation, (Beijing Jiaotong University), Ministry of Education · Weixin AI

Cross-Tokenizer Knowledge Distillation Language Model Distillation Knowledge Distillation

Feb 26, 2026

Probing to Refine: Reinforcement Distillation of LLMs via Explanatory Inversion

Distilling robust reasoning capabilities from large language models (LLMs) into smaller, computationally efficient student models remains an unresolved challenge. Despite recent advances, distilled models frequently suffer from superficial pattern memorization and subpar generalization. To overcome these limitations, we introduce a novel distillation framework that moves beyond simple mimicry to instill a deeper conceptual understanding. Our framework features two key innovations. First, to address pattern memorization, Explanatory Inversion (EI) generates targeted ``explanatory probes'' that compel the student to articulate the underlying logic behind an answer, rather than just memorizing it. Second, to improve generalization, Explanatory GRPO (EXGRPO) uses a reinforcement learning algorithm with a novel Dialogue Structure Utility Bonus, which explicitly rewards the student for maintaining a coherent reasoning process across these probes. Extensive evaluations on 12 datasets demonstrate significant improvements. Using Gemma-7b as the student model, our method yields an average 20.39% increase over zero-shot performance and a 6.02% improvement over the state-of-the-art distillation baselines. Moreover, models distilled with our method show remarkable training efficiency (e.g., surpassing vanilla fine-tuning with 10-25% training data) and strong generalization to out-of-distribution tasks. Implementation is released at https://github.com/Zhen-Tan-dmml/ExGRPO.git.

Zhen Tan, Chengshuai Zhao, Song Wang +3

Arizona State University · University of Virginia · University of North Carolina at Chapel Hill

Language Model Distillation Reasoning Distillation Knowledge Distillation RL for Language Model Reasoning Group Relative Policy Optimization

Feb 25, 2026

Decoder-based Sense Knowledge Distillation

Large language models (LLMs) learn contextual embeddings that capture rich semantic information, yet they often overlook structured lexical knowledge such as word senses and relationships. Prior work has shown that incorporating sense dictionaries can improve knowledge distillation for encoder models, but their application to decoder as generative models remains challenging. In this paper, we introduce Decoder-based Sense Knowledge Distillation (DSKD), a framework that integrates lexical resources into the training of decoder-style LLMs without requiring dictionary lookup at inference time. Extensive experiments on diverse benchmarks demonstrate that DSKD significantly enhances knowledge distillation performance for decoders, enabling generative models to inherit structured semantics while maintaining efficient training.

Qitong Wang, Mohammed J. Zaki, Georgios Kollias +1

Rensselaer Polytechnic Institute Troy, New York, USA · IBM Research Yorktown Heights, New York, USA

Language Model Distillation Knowledge Distillation Knowledge Augmentation for Language Models Decoder-Only Language Models

Feb 25, 2026

D-COT: Disciplined Chain-of-Thought Learning for Efficient Reasoning in Small Language Models

Chain-of-Thought (CoT) distillation from Large Language Models (LLMs) often induces "overthinking" in Small Language Models (SLMs), leading to performance degradation and excessive token consumption. In this study, we propose Disciplined Chain-of-Thought (D-CoT), a novel framework that enforces a structured reasoning process using control tags -- such as <TEMP_LOW> for fact-checking and <TEMP_HIGH> for multi-perspective exploration -- as auxiliary scaffolding during training. By optimizing the CoT trajectory, D-CoT suppresses reasoning drift and simultaneously achieves token reduction and performance improvement. We demonstrate the efficacy of our approach on Qwen3-8B: with only 5,000 training samples, D-CoT significantly boosts accuracy on GPQA-diamond by 9.9% and MMLU-Pro (0-shot) by 9.1%, while drastically reducing computational costs. Furthermore, we confirm that the model internalizes this disciplined thought structure, maintaining high performance even without explicit control tags during inference.

Shunsuke Ubukata

Toyo University

Efficient Language Model Reasoning CoT Distillation Language Model Distillation Overthinking in Language Models

Feb 25, 2026

Explain in Your Own Words: Improving Reasoning via Token-Selective Dual Knowledge Distillation

Knowledge Distillation (KD) can transfer the reasoning abilities of large models to smaller ones, which can reduce the costs to generate Chain-of-Thoughts for reasoning tasks. KD methods typically ask the student to mimic the teacher's distribution over the entire output. However, a student with limited capacity can be overwhelmed by such extensive supervision causing a distribution mismatch, especially in complex reasoning tasks. We propose Token-Selective Dual Knowledge Distillation (TSD-KD), a framework for student-centric distillation. TSD-KD focuses on distilling important tokens for reasoning and encourages the student to explain reasoning in its own words. TSD-KD combines indirect and direct distillation. Indirect distillation uses a weak form of feedback based on preference ranking. The student proposes candidate responses generated on its own; the teacher re-ranks those candidates as indirect feedback without enforcing its entire distribution. Direct distillation uses distribution matching; however, it selectively distills tokens based on the relative confidence between teacher and student. Finally, we add entropy regularization to maintain the student's confidence during distillation. Overall, our method provides the student with targeted and indirect feedback to support its own reasoning process and to facilitate self-improvement. The experiments show the state-of-the-art performance of TSD-KD on 10 challenging reasoning benchmarks, outperforming the baseline and runner-up in accuracy by up to 54.4% and 40.3%, respectively. Notably, a student trained by TSD-KD even outperformed its own teacher model in four cases by up to 20.3%. The source code is available at https://github.com/kmswin1/TSD-KD.

Minsang Kim, Seung Jun Baek

Korea University · SK Telecom

Language Model Distillation Selective Knowledge Distillation CoT Distillation LLM Reasoning

Feb 25, 2026

DWA-KD: Dual-Space Weighting and Time-Warped Alignment for Cross-Tokenizer Knowledge Distillation

Knowledge Distillation (KD) has emerged as a crucial technique for compressing Large Language Models (LLMs). Although existing cross-tokenizer KD methods have made notable progress, their effectiveness remains constrained by suboptimal alignment across sequence and vocabulary levels. To address these limitations, we introduce Dual-Space Weighting and Time-Warped Alignment (DWA-KD), a novel cross-tokenizer distillation framework that enhances token-wise distillation through dual-space entropy-based weighting and achieves precise sequence-level alignment by leveraging both lexical and semantic information. At the token level, DWA-KD maps teacher representations into the student space and vice versa, performing dual-space KD via Kullback-Leibler divergence (KL). The process is modulated by dual-space weights that up-weight tokens where the student is uncertain and the teacher is confident, thereby focusing learning on informative tokens rather than treating all positions equally. At the sequence level, DWA-KD applies Soft Dynamic Time Warping (Soft-DTW) to both the embedding and final hidden-state layers, enabling robust alignment of lexical and contextual semantics between teacher and student sequences. Extensive experiments across diverse NLP benchmarks demonstrate that DWA-KD outperforms state-of-the-art KD baselines, while ablation studies confirm the complementary contributions of entropy-based token weighting and embedding and final hidden state layer Soft-DTW alignment.

Duc Trung Vu, Pham Khanh Chi, Dat Phi Van +3

Hanoi University of Science and Technology, Vietnam · University of Monash, Australia

Cross-Tokenizer Knowledge Distillation Language Model Distillation Uncertainty-Aware Knowledge Distillation

Feb 24, 2026

Prompt-Level Distillation: A Non-Parametric Alternative to Model Fine-Tuning for Efficient Reasoning

Advanced reasoning typically requires Chain-of-Thought prompting, which is accurate but incurs prohibitive latency and substantial test-time inference costs. The standard alternative, fine-tuning smaller models, often sacrifices interpretability while introducing significant resource and operational overhead. To address these limitations, we introduce Prompt-Level Distillation (PLD). We extract explicit reasoning patterns from a Teacher model and organize them into a structured list of expressive instructions for the Student model's System Prompt. Evaluated using Gemma-3 4B, PLD improved Macro F1 scores on StereoSet (57% to 90.0%) and Contract-NLI (67% to 83%), while increasing LogiQA accuracy to 70%. Similar results on Mistral Small 3.1 demonstrate cross-architecture generalizability, enabling these compact models to match frontier performance with negligible latency overhead. These expressive instructions render the decision-making process transparent, allowing for full human verification of logic, making this approach ideal for regulated industries such as law, finance, and content moderation, as well as high-volume use cases and edge devices.

Sanket Badhe, Deep Shah

Google Mountain View, California, USA

Efficient Language Model Reasoning Language Model Distillation LLM Inference Acceleration

Feb 24, 2026

Don't Ignore the Tail: Decoupling top-K Probabilities for Efficient Language Model Distillation

The core learning signal used in language model distillation is the standard Kullback-Leibler (KL) divergence between the student and teacher distributions. Traditional KL divergence tends to be dominated by the next tokens with the highest probabilities, i.e., the teacher's modes, thereby diminishing the influence of less probable yet potentially informative components of the output distribution. We propose a new tail-aware divergence that decouples the contribution of the teacher model's top-K predicted probabilities from that of lower-probability predictions, while maintaining the same computational profile as the KL Divergence. Our decoupled approach reduces the impact of the teacher modes and, consequently, increases the contribution of the tail of the distribution. Experimental results demonstrate that our modified distillation method yields competitive performance in both pre-training and supervised distillation of decoder models across various datasets. Furthermore, the distillation process is efficient and can be performed with a modest academic budget for large datasets, eliminating the need for industry-scale computing.

Sayantan Dasgupta, Trevor Cohn, Timothy Baldwin

School of Computing and Information Systems, University of Melbourne, Melbourne, Australia.

Language Model Distillation Knowledge Distillation

Feb 18, 2026

Flow Map Language Models: One-step Language Modeling via Continuous Denoising

Language models based on discrete diffusion have attracted widespread interest for their potential to provide faster generation than autoregressive models. Despite their promise, these models typically produce samples whose quality sharply degrades in the few-step regime, preventing a dramatic speedup in practice. Here, we show that language models based on continuous flows over one-hot token embeddings can outperform discrete diffusion in both quality and speed. Importantly, our continuous formulation defines a unique flow map that can be learned directly for efficient few-step inference, a structure we show is unavailable to discrete methods. In this setting, we show that both the flow and its associated flow map can be learned with simple cross-entropy objectives that respect the simplex geometry of the data, and we identify three distinct choices for flow map distillation whose performance we compare in practice. Using these insights, we build a flow language model (FLM), a continuous flow that matches state-of-the-art discrete diffusion baselines on the One Billion Words (LM1B) and OpenWebText (OWT) datasets. We then distill FLM into a flow map language model (FMLM), whose one-step generation exceeds the 8-step quality of recent few-step discrete diffusion language models. Our work challenges the widely-held hypothesis that discrete noising processes are necessary for generative modeling over discrete modalities and paves the way toward accelerated language modeling at scale. Code is available at https://github.com/david3684/flm.

Chanhyuk Lee, Jaehoon Yoo, Manan Agarwal +6

KAIST · Carnegie Mellon University

Continuous Diffusion Language Models One-Step Generative Modeling Flow-Based Generative Modeling Generative Flow Networks Language Model Distillation

Feb 18, 2026

Aligning Language Models from User Interactions

Multi-turn user interactions are among the most abundant data produced by language models, yet we lack effective methods to learn from them. While typically discarded, these interactions often contain useful information: follow-up user messages may indicate that a response was incorrect, failed to follow an instruction, or did not align with the user's preferences. Importantly, language models are already able to make use of this information in context. After observing a user's follow-up, the same model is often able to revise its behavior. We leverage this ability to propose a principled and scalable method for learning directly from user interactions through self-distillation. By conditioning the model on the user's follow-up message and comparing the resulting token distribution with the original policy, we obtain a target for updating the policy that captures how the model's behavior changes in hindsight. We then distill this hindsight distribution back into the current policy. Remarkably, we show that training on real-world user conversations from WildChat improves language models across standard alignment and instruction-following benchmarks, without regressing other capabilities. The same mechanism enables personalization, allowing models to continually adapt to individual users through interaction without explicit feedback. Our results demonstrate that raw user interactions that arise naturally during deployment enable alignment, personalization, and continual adaptation.

Thomas Kleine Buening, Jonas Hübotter, Barna Pásztor +3

ETH Zurich · MIT · University of Zurich

LLM Alignment Personalized Language Model Alignment Self-Training for Language Models Continual Learning for LLMs Language Model Distillation

Feb 18, 2026

Task-Specific Knowledge Distillation via Intermediate Probes

Knowledge distillation from large language models (LLMs) assumes that the teacher's output distribution is a high-quality training signal. On reasoning tasks, this assumption is frequently violated. A model's intermediate representations may encode the correct answer, yet this information is lost or distorted through the vocabulary projection, where prompt formatting and answer-token choices creates brittle, noisy outputs. We introduce \method{}, a distillation framework that bypasses this bottleneck by training lightweight probes on frozen teacher hidden states and using the probe's predictions, rather than output logits, as supervision for student training. This simple change yields consistent improvements across four reasoning benchmarks (AQuA-RAT, ARC Easy/Challenge, and MMLU), with gains most pronounced under limited data. Probes trained on intermediate representations provide cleaner labels than the teacher's own outputs, effectively denoising the distillation signal. \method{} requires no architectural changes to student or teacher, is architecture-agnostic, and adds minimal compute since probe training is cheap and teacher representations can be cached. By exploiting internal representations, \method{} enables practitioners to extract more value from large teacher models without additional training data or architectural complexity.

Ryan Brown, Chris Russell

Oxford Internet Institute, University of Oxford, Oxford, United Kingdom.

Language Model Distillation Knowledge Distillation

Feb 16, 2026

Fast and Effective On-policy Distillation from Reasoning Prefixes

On-policy distillation (OPD), which samples trajectories from the student model and supervises them with a teacher at the token level, avoids relying solely on verifiable terminal rewards and can yield better generalization than off-policy distillation. However, OPD requires expensive on-the-fly sampling of the student policy during training, which substantially increases training cost, especially for long responses. Our initial analysis shows that, during OPD, training signals are often concentrated in the prefix of each output, and that even a short teacher-generated prefix can significantly help the student produce the correct answer. Motivated by these observations, we propose a simple yet effective modification of OPD: we apply the distillation objective only to prefixes of student-generated outputs and terminate each sampling early during distillation. Experiments on a suite of AI-for-Math and out-of-domain benchmarks show that on-policy prefix distillation matches the performance of full OPD while reducing training FLOP by 2x-47x.

Dongxu Zhang, Zhichao Yang, Sepehr Janghorbani +6

Optum AI

Language Model Distillation On-Policy Distillation

Feb 4, 2026

Semantic Self-Distillation for Language Model Uncertainty

Large language models present challenges for principled uncertainty quantification, in part due to their complexity and the diversity of their outputs. Semantic dispersion, or the variance in the meaning of sampled answers, has been proposed as a useful proxy for model uncertainty, but the associated computational cost prohibits its use in latency-critical applications. We show that sampled semantic distributions can be distilled into lightweight student models which estimate a prompt-conditioned density before the language model generates an answer token. The student model predicts a semantic distribution over possible answers; the entropy of this distribution provides a prompt-level uncertainty signal, and the probability density allows answer-level reliability evaluation. Across experiments on TriviaQA and MMLU, we find our student models perform competitively relative to the teacher's sampled semantic dispersion on a hallucination prediction task, whilst offering additional uncertainty primitives for out-of-domain detection and multiple-choice answer selection. We term this technique Semantic Self-Distillation (SSD), which can serve as a general framework for distilling predictive uncertainty in complex output spaces beyond language.

Edward Phillips, Sean Wu, Fredrik K. Gustafsson +2

Department of Engineering Science, University of Oxford · Oxford Suzhou Centre for Advanced Research, University of Oxford, Suzhou

Uncertainty Quantification Confidence Estimation in Language Models LLM Reliability Language Model Distillation LLM Uncertainty Estimation Semantic Uncertainty Estimation

Feb 3, 2026

Distilling LLM Reasoning into Graph of Concept Predictors

Deploying Large Language Models (LLMs) for discriminative workloads is often limited by inference latency, compute, and API costs at scale. Active distillation reduces these costs by querying an LLM oracle to train small discriminative students, but most pipelines distill only final labels, discarding intermediate reasoning signals and offering limited diagnostics of what reasoning is missing and where errors arise. We propose Graph of Concept Predictors (GCP), a reasoning-aware active distillation framework in which the teacher's reasoning is elicited as a directed acyclic graph of intermediate concepts and mirrored in the student. GCP enhances sample efficiency through a graph-aware acquisition strategy that weights per-concept uncertainty, gradient diversity, and coverage by node centrality. Additionally, it improves training stability and efficiency by performing targeted sub-module retraining, which attributes downstream loss to specific concept predictors and updates only the most influential modules. Experiments on eight NLP classification benchmarks demonstrate that GCP enhances performance under limited annotation budgets while yielding more interpretable and controllable training dynamics. Code is available at https://github.com/Ziyang-Yu/GCP.

Ziyang Yu, Liang Zhao

Department of Computer Science Emory University Atlanta, USA

CoT Distillation Active Learning Text Classification Language Model Distillation