On-Policy Context Distillation for Language Models

• URL: http://arxiv.org/abs/2602.12275v1
• Date: Thu, 12 Feb 2026 18:58:28 GMT
• Title: On-Policy Context Distillation for Language Models
• Authors: Tianzhu Ye, Li Dong, Xun Wu, Shaohan Huang, Furu Wei,
• Abstract summary: We propose On-Policy Context Distillation (OPCD), a framework that bridges on-policy distillation with context distillation.<n>We demonstrate the effectiveness of OPCD on two important applications: experiential knowledge distillation and system prompt distillation.
• Score: 92.82835176360864
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Context distillation enables language models to internalize in-context knowledge into their parameters. In our work, we propose On-Policy Context Distillation (OPCD), a framework that bridges on-policy distillation with context distillation by training a student model on its own generated trajectories while minimizing reverse Kullback-Leibler divergence against a context-conditioned teacher. We demonstrate the effectiveness of OPCD on two important applications: experiential knowledge distillation, where models extract and consolidate transferable knowledge from their historical solution traces, and system prompt distillation, where models internalize beneficial behaviors encoded in optimized prompts. Across mathematical reasoning, text-based games, and domain-specific tasks, OPCD consistently outperforms baseline methods, achieving higher task accuracy while better preserving out-of-distribution capabilities. We further show that OPCD enables effective cross-size distillation, where smaller student models can internalize experiential knowledge from larger teachers.

Related papers

• OVD: On-policy Verbal Distillation [47.727229201069555]
  On-policy Verbal Distillation (OVD) is a memory-efficient framework that replaces token-level probability matching with trajectory matching.<n>OVD dramatically reduces memory consumption while enabling on-policy distillation from teacher models with verbal feedback.
  arXiv  Detail & Related papers  (2026-01-29T16:48:14Z)
• Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models [44.041109669153506]
  On-Policy Self-Distillation (OPSD) is a framework where a single model acts as both teacher and student by conditioning on different contexts.<n>We demonstrate the efficacy of our method on multiple mathematical reasoning benchmarks.
  arXiv  Detail & Related papers  (2026-01-26T17:56:50Z)
• Honey, I Shrunk the Language Model: Impact of Knowledge Distillation Methods on Performance and Explainability [3.224880576815583]
  High computational and storage demands of Large Language Models limit their deployment in resource-constrained environments.<n>Previous research has introduced several distillation methods for both generating training data and for training the student model.<n>Despite their relevance, the effects of state-of-the-art distillation methods on model performance and explainability have not been thoroughly investigated.
  arXiv  Detail & Related papers  (2025-04-22T17:32:48Z)
• Learning from Stochastic Teacher Representations Using Student-Guided Knowledge Distillation [64.15918654558816]
  Self-distillation (SSD) training strategy is introduced for filtering and weighting teacher representation to distill from task-relevant representations only.<n> Experimental results on real-world affective computing, wearable/biosignal datasets from the UCR Archive, the HAR dataset, and image classification datasets show that the proposed SSD method can outperform state-of-the-art methods.
  arXiv  Detail & Related papers  (2025-04-19T14:08:56Z)
• TAID: Temporally Adaptive Interpolated Distillation for Efficient Knowledge Transfer in Language Models [6.8298782282181865]
  We introduce $textitTemporally Adaptive Interpolated Distillation (TAID)$, a novel knowledge distillation approach.<n>We show TAID's superior performance across various model sizes and architectures in both instruction tuning and pre-training scenarios.<n>These results demonstrate TAID's effectiveness in creating high-performing and efficient models, advancing the development of more accessible AI technologies.
  arXiv  Detail & Related papers  (2025-01-28T13:31:18Z)
• Learning to Maximize Mutual Information for Chain-of-Thought Distillation [13.660167848386806]
  Distilling Step-by-Step(DSS) has demonstrated promise by imbuing smaller models with the superior reasoning capabilities of their larger counterparts. However, DSS overlooks the intrinsic relationship between the two training tasks, leading to ineffective integration of CoT knowledge with the task of label prediction. We propose a variational approach to solve this problem using a learning-based method.
  arXiv  Detail & Related papers  (2024-03-05T22:21:45Z)
• Distilling Privileged Multimodal Information for Expression Recognition using Optimal Transport [46.91791643660991]
  Deep learning models for multimodal expression recognition have reached remarkable performance in controlled laboratory environments. These models struggle in the wild because of the unavailability and quality of modalities used for training. In practice, only a subset of the training-time modalities may be available at test time. Learning with privileged information enables models to exploit data from additional modalities that are only available during training.
  arXiv  Detail & Related papers  (2024-01-27T19:44:15Z)
• RAVEN: In-Context Learning with Retrieval-Augmented Encoder-Decoder Language Models [57.12888828853409]
  RAVEN is a model that combines retrieval-augmented masked language modeling and prefix language modeling. Fusion-in-Context Learning enables the model to leverage more in-context examples without requiring additional training. Our work underscores the potential of retrieval-augmented encoder-decoder language models for in-context learning.
  arXiv  Detail & Related papers  (2023-08-15T17:59:18Z)
• EmbedDistill: A Geometric Knowledge Distillation for Information Retrieval [83.79667141681418]
  Large neural models (such as Transformers) achieve state-of-the-art performance for information retrieval (IR) We propose a novel distillation approach that leverages the relative geometry among queries and documents learned by the large teacher model. We show that our approach successfully distills from both dual-encoder (DE) and cross-encoder (CE) teacher models to 1/10th size asymmetric students that can retain 95-97% of the teacher performance.
  arXiv  Detail & Related papers  (2023-01-27T22:04:37Z)
• Large Language Models with Controllable Working Memory [64.71038763708161]
  Large language models (LLMs) have led to a series of breakthroughs in natural language processing (NLP) What further sets these models apart is the massive amounts of world knowledge they internalize during pretraining. How the model's world knowledge interacts with the factual information presented in the context remains under explored.
  arXiv  Detail & Related papers  (2022-11-09T18:58:29Z)
• Collective Wisdom: Improving Low-resource Neural Machine Translation using Adaptive Knowledge Distillation [42.38435539241788]
  Scarcity of parallel sentence-pairs poses a significant hurdle for training high-quality Neural Machine Translation (NMT) models in bilingually low-resource scenarios. We propose an adaptive knowledge distillation approach to dynamically adjust the contribution of the teacher models during the distillation process. Experiments on transferring from a collection of six language pairs from IWSLT to five low-resource language-pairs from TED Talks demonstrate the effectiveness of our approach.
  arXiv  Detail & Related papers  (2020-10-12T04:26:46Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.