Learning to Correct: Calibrated Reinforcement Learning for Multi-Attempt Chain-of-Thought

arXiv:2604.17912v1 Announce Type: new Abstract: State-of-the-art reasoning models utilize long chain-of-thought (CoT) to solve increasingly complex problems using more test-time computation. In this work, we explore a long CoT setting where the model makes up to K successive attempts at solving a problem, in which each attempt is allowed to build on earlier ones after the model receives a hard verifier feedback. This motivates RL methods that can harness per-attempt rewards by carefully weighting individual attempts. We study optimizing the Verification@K reward (the model succeeds by the K-th attempt) and show that naively weighing the attempts by their pass/fail results in biased gradients. We introduce Calibrated Attempt-Level (CAL) GRPO by devising a weighing strategy to obtain unbiased gradients while maintaining small variance. Our theory reveals how incorporating per-attempt rewards influence the training and the eventual Verification@K performance. Experiments, baselines, and ablations on synthetic and real data corroborate our theory and the benefits of CAL-GRPO over vanilla GRPO as well as naive weighting.

Related

• What If Consensus Lies? Selective-Complementary Reinforcement Learning at Test Time
• Self-Aligned Reward: Towards Effective and Efficient Reasoners
• SCATR: Simple Calibrated Test-Time Ranking
• Stability and Generalization in Looped Transformers
• The Illusion of Superposition? A Principled Analysis of Latent Thinking in Language Models
• On the Step Length Confounding in LLM Reasoning Data Selection

Source: arXiv cs.LG | 2026-04-21