English
Related papers

Related papers: Same Evidence, Different Answers: Canonical-Contex…

200 papers

On-policy distillation (OPD) has recently emerged as an effective post-training paradigm for consolidating the capabilities of specialized expert models into a single student model. Despite its empirical success, the conditions under which…

On-policy distillation (OPD) is increasingly used in LLM post-training because it can leverage a teacher model to provide dense supervision on student rollouts. The standard implementation, however, usually reduces distribution matching to…

Machine Learning · Computer Science 2026-04-28 Yuqian Fu , Haohuan Huang , Kaiwen Jiang , Jiacai Liu , Zhuo Jiang , Yuanheng Zhu , Dongbin Zhao

Black-box knowledge distillation for large language models presents a strict trade-off. Simple off-policy methods (e.g., sequence-level knowledge distillation) struggle to correct the student's inherent errors. Fully on-policy methods…

Machine Learning · Computer Science 2026-04-24 Xiwen Chen , Jingjing Wang , Wenhui Zhu , Peijie Qiu , Xuanzhao Dong , Hejian Sang , Zhipeng Wang , Alborz Geramifard , Feng Luo

Large language models often solve tasks from a fully specified prompt but degrade when the same requirements unfold over multiple turns, known as the lost-in-conversation (LiC) gap. We trace part of this degradation to self-contamination:…

Computation and Language · Computer Science 2026-05-27 Haoyu Zheng , Yun Zhu , Shu Yuan , Shangming Chen , Qing Wang , Wenqiao Zhang , Jun Xiao , Yueting Zhuang

In-context learning (ICL) allows large language models (LLMs) to solve novel tasks without weight updates. Despite its empirical success, the mechanism behind ICL remains poorly understood, limiting our ability to interpret, improve, and…

Machine Learning · Computer Science 2025-06-16 Chengye Li , Haiyun Liu , Yuanxi Li

Recently, multi-modal content generation has attracted lots of attention from researchers by investigating the utilization of visual instruction tuning based on large language models (LLMs). To enhance the performance and generalization…

Computer Vision and Pattern Recognition · Computer Science 2023-11-15 Xinwei Li , Li Lin , Shuai Wang , Chen Qian

Large language models (LMs) beyond a certain scale, demonstrate the emergent capability of generating free-text rationales for their predictions via chain-of-thought (CoT) prompting. While CoT can yield dramatically improved performance,…

Computation and Language · Computer Science 2023-09-01 Peifeng Wang , Zhengyang Wang , Zheng Li , Yifan Gao , Bing Yin , Xiang Ren

GRPO-style RLVR trains reasoning models from multiple on-policy attempts per prompt, but typically uses these attempts only through terminal rewards. We show that a mixed group contains a richer process signal: a correct completion is a…

Machine Learning · Computer Science 2026-05-19 Zhiquan Tan , Yinrong Hong

On-Policy Distillation (OPD) has gained wide attraction as an LLM post-training paradigm due to its effectiveness in improving capabilities without introducing model distribution drift, and consequently, regression in general tasks.…

Artificial Intelligence · Computer Science 2026-05-25 Aristotelis Lazaridis , Dylan Bates , Aman Sharma , Brian King , Vincent Lu , Jack FitzGerald

On-Policy Self-Distillation (OPSD) is a unified learning framework in which a single large language model acts simultaneously as both teacher and student. Unlike conventional knowledge distillation that relies on a separate, often larger…

Human-Computer Interaction · Computer Science 2026-05-22 Fangming Cui , Sunan Li , Jiahong Li

On-policy distillation (OPD) trains a student model on its own rollouts using dense feedback from a stronger teacher. Prior literature suggests that, provided teacher feedback is available, supervising the full sequence of response tokens…

Computation and Language · Computer Science 2026-05-26 Kaiyuan Liu , Ziyuan Zhuang , Yang Bai , Bing Wang , Rongxiang Weng , Jieping Ye

Language models significantly benefit from context tokens, such as prompts or scratchpads. They perform better when prompted with informative instructions, and they acquire new reasoning capabilities by generating a scratch-pad before…

Computation and Language · Computer Science 2022-10-03 Charlie Snell , Dan Klein , Ruiqi Zhong

On-policy reinforcement learning has become the dominant paradigm for reasoning alignment in large language models, yet its sparse, outcome-level rewards make token-level credit assignment notoriously difficult. On-Policy Distillation (OPD)…

Machine Learning · Computer Science 2026-04-14 Binbin Zheng , Xing Ma , Yiheng Liang , Jingqing Ruan , Xiaoliang Fu , Kepeng Lin , Benchang Zhu , Ke Zeng , Xunliang Cai

On-Policy Distillation (OPD) has emerged as a dominant post-training paradigm for large language models, especially for reasoning domains. However, OPD remains unstable in practice due to the high gradient variance of its single-sample…

Machine Learning · Computer Science 2026-05-11 Minjae Oh , Sangjun Song , Gyubin Choi , Yunho Choi , Yohan Jo

Proprietary large language models (LLMs) embody substantial economic value and are generally exposed only as black-box APIs, yet adversaries can still exploit their outputs to extract knowledge via distillation. Existing defenses focus…

Computation and Language · Computer Science 2026-05-07 Hao Fang , Tianyi Zhang , Tianqu Zhuang , Jiawei Kong , Kuofeng Gao , Bin Chen , Leqi Zheng , Shu-Tao Xia , Ke Xu

Large language models often struggle to recognize their knowledge limits in closed-book question answering, leading to confident hallucinations. While decomposed prompting is typically used to improve accuracy, we investigate its impact on…

Computation and Language · Computer Science 2026-02-05 Dhruv Madhwal , Lyuxin David Zhang , Dan Roth , Tomer Wolfson , Vivek Gupta

Improving large language model (LLM) reasoning requires supervision that is both aligned with the model's own test-time states and informative at the token level. Reinforcement learning with verifiable rewards provides on-policy exploration…

Machine Learning · Computer Science 2026-04-30 Zhiquan Tan , Yinrong Hong

The advent of contextual word embeddings -- representations of words which incorporate semantic and syntactic information from their context -- has led to tremendous improvements on a wide variety of NLP tasks. However, recent contextual…

Computation and Language · Computer Science 2021-06-09 Prakhar Gupta , Martin Jaggi

While the shift from cascaded dialogue systems to end-to-end (E2E) speech Large Language Models (LLMs) improves latency and paralinguistic modeling, E2E models often exhibit a significant performance degradation compared to their text-based…

Audio and Speech Processing · Electrical Eng. & Systems 2026-03-31 Di Cao , Dongjie Fu , Hai Yu , Siqi Zheng , Xu Tan , Tao Jin

Large language models (LLMs) have showcased remarkable capabilities in complex reasoning through chain of thought (CoT) prompting. Recently, there has been a growing interest in transferring these reasoning abilities from LLMs to smaller…

Computation and Language · Computer Science 2023-12-21 Hongzhan Chen , Siyue Wu , Xiaojun Quan , Rui Wang , Ming Yan , Ji Zhang