10 papers
Revealing Behavioral Plasticity in Large Language Models: A Token-Conditional Perspective
Liyuan Mao, Le Yu, Jing Zhou +7
In this work, we reveal that Large Language Models (LLMs) possess intrinsic behavioral plasticity-akin to chameleons adapting their coloration to environmental cues-that can be exp…
Stealth Fine-Tuning: Efficiently Breaking Alignment in RVLMs Using Self-Generated CoT
Le Yu, Zhengyue Zhao, Yawen Zheng +1
Reasoning-augmented Vision-Language Models (RVLMs) rely on safety alignment to prevent harmful behavior, yet their exposed chain-of-thought (CoT) traces introduce new attack surfac…
Outcome Accuracy is Not Enough: Aligning the Reasoning Process of Reward Models
Binghai Wang, Yantao Liu, Yuxuan Liu +13
Generative Reward Models (GenRMs) and LLM-as-a-Judge exhibit deceptive alignment by producing correct judgments for incorrect reasons, as they are trained and evaluated to prioriti…
Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
Shenzhi Wang, Le Yu, Chang Gao +15
Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a powerful approach to enhancing the reasoning capabilities of Large Language Models (LLMs), while its mechanis…
Beyond Turn Limits: Training Deep Search Agents with Dynamic Context Window
Qiaoyu Tang, Hao Xiang, Le Yu +11
While recent advances in reasoning models have demonstrated cognitive behaviors through reinforcement learning, existing approaches struggle to invoke deep reasoning capabilities i…
RefCritic: Training Long Chain-of-Thought Critic Models with Refinement Feedback
Qiaoyu Tang, Hao Xiang, Le Yu +6
With the rapid advancement of Large Language Models (LLMs), developing effective critic modules for precise guidance has become crucial yet challenging. In this paper, we initially…