1
Knowing When to Quit: A Principled Framework for Dynamic Abstention in LLM Reasoning
让LLM学会"知难而退"：一个原则性框架，帮助模型动态判断何时应该放弃推理，提升可靠性与效率。
arXiv:2604.18419v4 Announce Type: replace Abstract: LLMs utilizing chain-of-thought reasoning often waste substantial compute by producing long, incor…