12 citations · 60 across the 72 of their papers we have counts for
8 papers · 1 filter
A Token-Level Analysis of Sampled-Token Reverse-KL On-Policy Distillation
Bing Shao, Jiazheng Zhang, Long Ma +12
On-policy distillation (OPD) supervises a student on its own trajectories with token-level signals from a frozen teacher, yet how a sampled loss allocates updates across tokens rem…
From Parameters to Data: A Task-Parameter-Guided Fine-Tuning Pipeline for Efficient LLM Alignment
Hao Chen, Qi Zhang, Liyao Li +7
Adapting Large Language Models (LLMs) to specialized domains typically incurs high data and computational overhead. While prior efficiency efforts have largely treated data selecti…
Entropy Polarity in Reinforcement Fine-Tuning: Direction, Asymmetry, and Control
Jiazheng Zhang, Ziche Fu, Junrui Shen +17
Policy entropy has emerged as a fundamental measure for understanding and controlling exploration in reinforcement learning with verifiable rewards (RLVR) for LLMs. However, existi…
Improving RL Exploration for LLM Reasoning through Retrospective Replay
Shihan Dou, Muling Wu, Jingwen Xu +4
Reinforcement learning (RL) has increasingly become a pivotal technique in the post-training of large language models (LLMs). The effective exploration of the output space is essen…
MetaRM: Shifted Distributions Alignment via Meta-Learning
Shihan Dou, Yan Liu, Enyu Zhou +9
The success of Reinforcement Learning from Human Feedback (RLHF) in language model alignment is critically dependent on the capability of the reward model (RM). However, as the tra…
Subspace Defense: Discarding Adversarial Perturbations by Learning a Subspace for Clean Signals
Rui Zheng, Yuhao Zhou, Zhiheng Xi +3
Deep neural networks (DNNs) are notoriously vulnerable to adversarial attacks that place carefully crafted perturbations on normal examples to fool DNNs. To better understand such…
Making Harmful Behaviors Unlearnable for Large Language Models
Xin Zhou, Yi Lu, Ruotian Ma +3
Large language models (LLMs) have shown great potential as general-purpose AI assistants in various domains. To meet the requirements of different applications, LLMs are often cust…
Improving Generalization of Alignment with Human Preferences through Group Invariant Learning
Rui Zheng, Wei Shen, Yuan Hua +9
The success of AI assistants based on language models (LLMs) hinges crucially on Reinforcement Learning from Human Feedback (RLHF), which enables the generation of responses more a…