Surprising Success, Repeated Failure: Entropy-Guided Credit Assignment for Exploration in LLM Reasoning

Benchmark Model Rank Results
mathematical-reasoning-on-hmmt-feb-2026Olmo-3-7B-Think-DPO + EAPO#10Accuracy: 45.5
mathematical-reasoning-on-aime24Olmo-3-7B-Think-DPO + EAPO#39Acc: 76.7
mathematical-reasoning-on-aime-2026Olmo-3-7B-Think-DPO + EAPO#41Accuracy: 70.3
mathematical-reasoning-on-aime-2025Olmo-3-7B-Think-DPO + EAPO#62Accuracy: 65.3