发现大模型藏有未被使用的正确知识，能提升问答准确率。
Are LLMs Really Not Knowledgeable? Mining the Submerged Knowledge in LLMs' Memory
- 通过分析输出概率分布，挖掘模型内部隐藏的正确答案
- 发现模型实际掌握的知识远超标准问答准确率显示的水平
- 提醒提示词设计别压制低置信度但正确的回答
大语言模型虽常在问答任务中因幻觉和不确定而表现不佳，但研究发现其参数中仍保留大量正确知识。通过分析标记级输出分布，我们发现正确答案常出现在高概率候选中，却未被选中。为此提出新指标Hits@k，评估模型隐含的知识保留能力。实验表明，模型的实际事实知识远超传统问答准确率反映的程度。进一步分析发现，允许生成“不确定”答案的少样本提示策略会无意中抑制正确答案的出现。通过定量实验量化了这种抑制效应，为知识密集型任务中的提示与解码设计提供实用指导。
原文摘要 · Abstract (English)
Large language models (LLMs) have shown promise as parametric knowledge bases, but often underperform on question answering (QA) tasks due to hallucinations and uncertainty. While prior work attributes these failures to knowledge gaps in the model's parameters, we uncover a complementary phenomenon: LLMs frequently retain correct knowledge even when generating incorrect or "unsure" answers. By analyzing the token-level output distributions, we find that correct answers often appear among high-probability candidates, despite not being selected. Motivated by this, we propose Hits@k, a novel metric to evaluate latent knowledge retention independent of answer surface form. Our experiments reveal that LLMs possess significantly more factual knowledge than is reflected by standard QA accuracy. Building on these insights, we further examine the prevailing few-shot QA paradigm. We find that prompting strategies which allow "unsure" outputs can inadvertently suppress correct answers by discouraging low-confidence generation. We design a set of quantitative experiments to measure this suppression effect, offering practical guidance for future prompt and decoding design in knowledge-intensive tasks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。