Knowing When to Quit: A Principled Framework for Dynamic Abstention in LLM Reasoning 事件

Knowing When to Quit: A Principled Framework for Dynamic Abstention in LLM Reasoning arXiv:2604.18419v4 Announce Type: replace-cross Abstract: LLMs utilizing chain-of-thought reasoning often waste substantial compute by producing long, incorrect responses. Abstention can mitigate this by withholding outputs unlikely to be correct. While most abstention methods decide to withhold outputs before or after generation, dynamic mid-generation abstention considers early termination of unpromising reas