← Back to all news
arXiv AI By Xun Lu, Shumeng Yang, Xinrui Wang, Songjun Tu, Linjing Li, Daniel Dajun Zeng

Learning to Reason with Selective Internal Feedback

Read the original on arXiv AI →

The paper introduces SIF, a reinforcement learning technique that uses selective internal feedback to improve language models’ reasoning without external verifiers. SIF constructs comparative rewards by comparing token‑level self‑certainty across multiple windows and peers, rewarding only when both voting levels agree, and applies sample and group‑level filtering to decide which responses influence policy updates. Experiments on three backbone models across 12 mathematical and scientific benchmarks show that SIF’s final checkpoints outperform the best‑validation checkpoints of Intuitor and Vigor, achieving a 4.71‑point accuracy gain on Qwen2.5‑Math‑1.5B and maintaining accuracy during continued optimization with minimal test‑time rollouts.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

Related stories

arXiv AI
Sep 28

Learning to Stop without Learning to Stop: Self-Supervised Confidence Training Improves Reasoning Efficiency

The paper demonstrates that fine‑tuning reasoning models to predict their own confidence at intermediate steps—using only 600 self‑supervised examples—substantially improves inference efficiency. Without adding any explicit stopping or length penalties, the models generate up to 25 % fewer tokens while maintaining accuracy on mathematical, scientific, and coding benchmarks across several architectures. The study finds that confidence supervision preserves the models’ high‑level reasoning structure rather than merely suppressing specific behaviors.

By Parsa Hosseini, Akasha Tigalappanavara, Sumit Nawathe, Chenrui Fan, Sourya Basu, Genta Indra Winata, Anirban Das, Soheil Feizi, Nima Chitsazan
llmsreinforcement-learningfine-tuningbenchmarks
More like this →
arXiv Machine Learning
Jun 9

ConSteer-RL: Steering Reasoning Capabilities in Large Language Models via Confidence-Aware Reinforcement Learning

arXiv:2606. 08088v1 Announce Type: new Abstract: Reinforcement Learning from Verifiable Rewards (RLVR) has recently become a key paradigm for improving the reasoning abilities of Large Language Models (LLMs), yet it remains limited by sparse binary rewards and its ignorance of model-internal uncertainty.

By Qing Miao, Yiming Zhao, Jing Yang, Chenxi Liu, Yuehai Chen, Yuewen Liu, Shaoyi Du, Badong Chen
llmsreinforcement-learning
More like this →