First Return, Entropy-Eliciting Explore

TY.Zheng, xts, Qingshui Gu, Taoran Liang, Xingwei Qu, Xin Zhou, Yizhi Li, Maxwell Wen, Chenghua Lin, Wenhao Huang, Qian Liu, Ge Zhang, Zejun Ma

First Return, Entropy-Eliciting Explore: 24 upvotes on Hugging Face Daily Papers, #8 of 22 papers on 2025-07-10. Day-by-day upvote history.

Reinforcement Learning from Verifiable Rewards (RLVR) improves the reasoning abilities of Large Language Models (LLMs) but it struggles with unstable exploration. We propose FR3E (First Return, Entropy-Eliciting Explore), a structured exploration framework that identifies high-uncertainty decision points in reasoning trajectories and performs targeted rollouts to construct semantically grounded intermediate feedback. Our method provides targeted guidance without relying on dense supervision. Empirical results on mathematical reasoning benchmarks(AIME24) show that FR3E promotes more stable training, produces longer and more coherent responses, and increases the proportion of fully correct trajectories. These results highlight the framework's effectiveness in improving LLM reasoning through more robust and structured exploration.

Paper page on Hugging Face · arXiv

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.