让大模型自己说出信心分数，简单有效评估回答可靠性。
On Verbalized Confidence Scores for LLMs
- 让模型在输出中直接生成信心分数，无需额外模型或复杂计算。
- 不同提示方法下信心分数的可靠性差异显著，部分方法可实现良好校准。
- 适合希望低成本提升模型可信度的开发者与研究者使用。
大语言模型（LLMs）日益融入日常生活，其可信度至关重要。量化不确定性有助于增强人类对模型输出的信任，并使模型代理基于彼此的不确定性做出更明智决策。现有方法包括利用内部词元概率、特定任务代理模型或多次采样。本文提出让模型自身以自然语言形式输出信心分数，实现无需提示和模型依赖的低开销不确定性量化。通过大规模基准测试，我们评估了不同数据集、模型和提示方式下该方法的可靠性。结果表明，信心分数的可靠性高度依赖于提问方式，但某些提示策略可生成校准良好的分数。我们认为，这种自我表述的信心评分有望成为未来简单而通用的不确定性量化方法。代码已公开：https://github.com/danielyxyang/llm-verbalized-uq。
原文摘要 · Abstract (English)
The rise of large language models (LLMs) and their tight integration into our daily life make it essential to dedicate efforts towards their trustworthiness. Uncertainty quantification for LLMs can establish more human trust into their responses, but also allows LLM agents to make more informed decisions based on each other's uncertainty. To estimate the uncertainty in a response, internal token logits, task-specific proxy models, or sampling of multiple responses are commonly used. This work focuses on asking the LLM itself to verbalize its uncertainty with a confidence score as part of its output tokens, which is a promising way for prompt- and model-agnostic uncertainty quantification with low overhead. Using an extensive benchmark, we assess the reliability of verbalized confidence scores with respect to different datasets, models, and prompt methods. Our results reveal that the reliability of these scores strongly depends on how the model is asked, but also that it is possible to extract well-calibrated confidence scores with certain prompt methods. We argue that verbalized confidence scores can become a simple but effective and versatile uncertainty quantification method in the future. Our code is available at https://github.com/danielyxyang/llm-verbalized-uq.
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。