paper-with-me

Papers

HEALing Entropy Collapse: Enhancing Exploration in Few-Shot RLVR via Hybrid-Domain Entropy Dynamics Alignment

2026-04-20 · Zhanyu Liu, Qingguo Hu, Ante Wang, Chenqing Liu, Zhishang Xiang, Hui Li, Delai Qiu, Jinsong Su arxiv

Reinforcement Learning with Verifiable Reward (RLVR) has proven effective for training reasoning-oriented large language models, but existing methods largely assume high-resource settings with abundant training data. In low-resource scenarios, RLVR is prone to more severe entropy collapse, which substantially limits exploration and degrades reasoning performance. To address this issue, we propose Hybrid-domain Entropy dynamics ALignment (HEAL), a framework tailored for few-shot RLVR. HEAL first selectively incorporates high-value general-domain data to promote more diverse exploration. Then, we introduce Entropy Dynamics Alignment (EDA), a reward mechanism that aligns trajectory-level entropy dynamics between the target and general domains, capturing both entropy magnitude and fine-grained variation. Through this alignment, EDA not only further mitigates entropy collapse but also encourages the policy to acquire more diverse exploration behaviors from the general domain. Experiments across multiple domains show that HEAL consistently improves few-shot RLVR performance. Notably, using only 32 target-domain samples, HEAL matches or even surpasses full-shot RLVR trained with 1K target-domain samples.

📄 PDF Abstract BibTeX arXiv:2604.17928

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning

Similar Papers 제목 키워드 기반

Clip-Low Increases Entropy and Clip-High Decreases Entropy in Reinforcement Learning of Large Language Models

2025-09-30 · Jaesung R. Park, Junsu Kim, Gyeongman Kim, Jinyoung Jo 외 arxiv

Reinforcement learning with verifiable rewards (RLVR) has recently emerged as the leading approach for enhancing the reasoning capabilities of large language models (LLMs). However, RLVR is prone to entropy collapse, whe…

Reinforcement Learning

Efficient Reinforcement Learning with Semantic and Token Entropy for LLM Reasoning

2025-12-04 · Hongye Cao, Zhixin Bai, Ziyue Peng, Boyan Wang 외 arxiv

Reinforcement learning with verifiable rewards (RLVR) has demonstrated superior performance in enhancing the reasoning capability of large language models (LLMs). However, this accuracy-oriented learning paradigm often s…

Reinforcement Learning

Shattering the Autoregressive Curse: Dynamic Epistemic Entropy Orchestrated Erasable Reinforcement Learning for LLMs

2026-06-16 · Ziliang Wang, Kang An, Faqiang Qian, Jialu Cai 외 arxiv

Although reinforcement learning (RL) has expanded the cognitive boundaries of large language models (LLMs), it often remains vulnerable to the autoregressive curse in long-horizon logical reasoning: small epistemic pertu…

Reinforcement LearningMathematical ReasoningLogical Reasoning

CURE: Critical-Token-Guided Re-Concatenation for Entropy-Collapse Prevention

2025-08-14 · Qingbin Li, Rongkun Xue, Jie Wang, Ming Zhou 외 arxiv

Recent advances in Reinforcement Learning with Verified Reward (RLVR) have driven the emergence of more sophisticated cognitive behaviors in large language models (LLMs), thereby enhancing their reasoning capabilities. H…

Reinforcement Learning

The Road Less Traveled: Enhancing Exploration in LLMs via Sequential Sampling

2025-10-17 · Shijia Kang, Muhan Zhang arxiv

Reinforcement learning (RL) has been pivotal in enhancing the reasoning capabilities of large language models (LLMs), but it often suffers from limited exploration and entropy collapse, where models exploit a narrow set …

Reinforcement Learning