8 papers
Distill Where You Fail: Recovering Learning Signals of Negative RL-Groups from Adaptive Teacher Guidance
Zhuowen Han, Jinwei Xiao, Zhengxi Lu +9
Reinforcement learning with verifiable rewards (RLVR) has become a standard paradigm for post-training large language models (LLMs). While Group Relative Policy Optimization (GRPO)…
Thinking Seeds: Leveraging Historical Diversity for Position-Aware RL in LLMs
Lei Yang, Wei Bi, Chenxi Sun +2
On-policy reinforcement learning (RL) for language model post-training suffers from a fundamental tension: as training progresses, policy entropy collapses and sampling diversity d…
From Insight to Action: A Novel Framework for Interpretability-Guided Data Selection in Large Language Models
Ling Shi, Xinwei Wu, Xiaohu Zhao +7
While mechanistic interpretability tools like Sparse Autoencoders (SAEs) can uncover meaningful features within Large Language Models (LLMs), a critical gap remains in transforming…
Why Does Reinforcement Learning Generalize? A Feature-Level Mechanistic Study of Post-Training in Large Language Models
Dan Shi, Zhuowen Han, Simon Ostermann +3
Reinforcement learning (RL)-based post-training often improves the reasoning performance of large language models (LLMs) beyond the training domain, while supervised fine-tuning (S…
Revisiting Entropy in Reinforcement Learning for Large Reasoning Models
Renren Jin, Pengzhi Gao, Yuqi Ren +6
Reinforcement learning with verifiable rewards (RLVR) has emerged as a prominent paradigm for enhancing the reasoning capabilities of large language models (LLMs). However, the ent…
TaP: A Taxonomy-Guided Framework for Automated and Scalable Preference Data Generation
Renren Jin, Tianhao Shen, Xinwei Wu +9
Conducting supervised and preference fine-tuning of large language models (LLMs) requires high-quality datasets to improve their ability to follow instructions and align with human…