用特定指令数据训练大模型承认无知，显著降低幻觉率。
Inducing Epistemological Humility in Large Language Models: A Targeted SFT Approach to Reducing Hallucination
- 设计含虚构术语的问题数据集，教会模型识别知识边界。
- 在多个模型上测试，幻觉检测得分提升最高达25.91%。
- 无需强化学习，适合想提升模型可信度的研究者。
大型语言模型常产生流畅但错误的信息，部分原因在于监督微调（SFT）隐式奖励始终作答。我们提出$ extit{HypoTermInstruct}$，一个包含31,487条回复、覆盖11,151个问题的SFT数据集，旨在让模型学会元认知谦逊——即识别自身知识局限并承认不确定性。方法通过关于不存在的“假设性”术语的问题实现。同时发布$ extit{HypoTermQA-Enhanced}$基准，通过多轮验证增强幻觉检测能力。我们在$ extit{Llama3.1-8B}$和$ extit{Gemma3-4B}$（base与instruct版）上执行800次受控LoRA SFT实验，测试100种微调配置并设置对照组。结果表明，用$ extit{HypoTermInstruct}$替代通用指令数据，可使HypoTerm Score中位数提升0.19%至25.91%，FactScore提升0.39%至0.86%，同时在MMLU上性能仅轻微下降0.26%至0.35%。研究证明，针对元认知技能的高质量目标SFT数据能有效减少幻觉，且无需偏好或强化学习流程，为构建更可靠的AI系统提供机制理解与实用路径。
原文摘要 · Abstract (English)
Large language models (LLMs) often hallucinate, producing fluent but false information, partly because supervised fine-tuning (SFT) implicitly rewards always responding. We introduce $\textit{HypoTermInstruct}$, an SFT dataset (31,487 responses for 11,151 questions) designed to teach models epistemological humility-the ability to recognize the limits of their own knowledge and admit uncertainty. This is achieved through questions about non-existent "hypothetical" terms. We also release $\textit{HypoTermQA-Enhanced}$, a benchmark for hallucination tendency strengthened through multiple validations. We conducted 800 controlled LoRA SFT runs across $\textit{Llama3.1-8B}$ and $\textit{Gemma3-4B}$ (base and instruct), testing 100 fine-tuning configurations with paired controls. Our results demonstrate that replacing generic instruction data with $\textit{HypoTermInstruct}$ significantly improves the HypoTerm Score (median increases of 0.19% to 25.91%) and FactScore (+0.39% to +0.86%), while maintaining stable performance on MMLU (minimal decreases of 0.26% to 0.35%). Our work demonstrates that targeted, high-quality SFT data teaching meta-cognitive skills can effectively reduce hallucination without preference/RL pipelines, providing mechanistic insights and a practical path toward more reliable AI systems.
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。