分解式提示无法弥补知识差距，但有助于模型说“我不知道”
Decomposed Prompting Does Not Fix Knowledge Gaps, But Helps Models Say "I Don't Know"
AI总结：
研究大语言模型在闭卷问答中识别知识局限的问题，评估三种提示方式在不同模型规模和多跳问答基准下的影响，利用提示方式间的分歧信号实现无训练弃权策略，提升模型可靠性。
摘要：
大语言模型在闭卷问答中难以识别知识局限，导致产生幻觉。虽分解式提示通常用于提高准确性，但本文研究其对可靠性的影响。评估了三种任务等效的提示方式，发现前沿模型中分解带来的准确性提升减弱，但提示方式间的分歧仍能指示潜在错误，据此实现无训练弃权策略，效果优于标准不确定性基线。
英文摘要：
Large language models often struggle to recognize their knowledge limits in closed-book question answering, leading to confident hallucinations. While decomposed prompting is typically used to improve accuracy, we investigate its impact on reliability. We evaluate three task-equivalent prompting regimes: Direct, Assistive, and Incremental, across different model scales and multi-hop QA benchmarks. We find that although accuracy gains from decomposition diminish in frontier models, disagreements between prompting regimes remain highly indicative of potential errors. Because factual knowledge is typically stable while hallucinations are stochastic, cross-regime agreement provides a precise signal of internal uncertainty. We leverage this signal to implement a training-free abstention policy that requires no retrieval or fine-tuning. Our results show that disagreement-based abstention outperforms standard uncertainty baselines as an error detector, improving both F1 and AUROC across settings. This demonstrates that decomposition-based prompting can serve as a practical diagnostic probe for model reliability in closed-book QA.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
