How Do Answer Tokens Read Reasoning Traces? Self-Reading Patterns in Thinking LLMs for Quantitative ReasoningMay 8, 2026·Haoyang Chen,Yi Liu,Jianzhi Shao,Tao Zhang,Chengfu HuoWei Hu· 0 min read Cite URLTypeConference paperPublicationFindings of the Association for Computational Linguistics: ACL 2026Last updated on May 8, 2026AuthorsWei HuNanjing University School of Computer Science ← Focusing Condition: Inference-Time Self-Contrastive Steering Elicits Better Conditional Text Embeddings in LLMs May 8, 2026PEGRL: Improving Machine Translation by Post-Editing Guided Reinforcement Learning May 8, 2026 →
