对比搜索法能更准确估计大模型输出的不确定性。
Decoding Uncertainty: The Impact of Decoding Strategies for Uncertainty Estimation in Large Language Models
- 用对比搜索法调整概率分布，减少重复生成。
- 在多种对齐模型中，该方法平均提升不确定性估计效果。
- 适用于关注生成可信度的研究者或部署场景。
解码策略会改变语言模型输出的概率分布，从而影响生成质量和不确定性估计。本研究探究了解码策略对大语言模型（LLMs）不确定性估计的影响。实验表明，对比搜索法（Contrastive Search）通过缓解重复问题，在多种偏好对齐的LLM上均能获得更优的不确定性估计。然而，当模型仅经过监督微调（无显式对齐）时，这些策略的优势可能不一致。
原文摘要 · Abstract (English)
Decoding strategies manipulate the probability distribution underlying the output of a language model and can therefore affect both generation quality and its uncertainty. In this study, we investigate the impact of decoding strategies on uncertainty estimation in Large Language Models (LLMs). Our experiments show that Contrastive Search, which mitigates repetition, yields better uncertainty estimates on average across a range of preference-aligned LLMs. In contrast, the benefits of these strategies sometimes diverge when the model is only post-trained with supervised fine-tuning, i.e. without explicit alignment.
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。