数据增强能提升大模型置信度估计，缓解过自信问题。
The Effects of Data Augmentation on Confidence Estimation for LLMs
- 用多种数据增强方法改进置信度估计。
- 增强后模型更少过自信，尤其在语义不变时多样性越高效果越好。
- 随机组合增强策略兼具实用性和可迁移性，适合实际应用。
置信度估计对反映大语言模型（LLMs）的可靠性至关重要，尤其在广泛使用的闭源模型中。尽管利用数据增强进行置信度估计是可行的，但现有研究多聚焦于特定增强技术，限制了其潜力。本文研究不同数据增强方法对置信度估计的影响。结果表明，数据增强策略能显著提升性能，并缓解过自信现象。我们进一步发现，在保持语义信息的前提下，更高的数据多样性可增强增强效果；不同增强策略的效果在不同应用场景中存在差异。考虑到参数可迁移性和使用便捷性，随机组合多种增强方法是一种有前景的选择。
原文摘要 · Abstract (English)
Confidence estimation is crucial for reflecting the reliability of large language models (LLMs), particularly in the widely used closed-source models. Utilizing data augmentation for confidence estimation is viable, but discussions focus on specific augmentation techniques, limiting its potential. We study the impact of different data augmentation methods on confidence estimation. Our findings indicate that data augmentation strategies can achieve better performance and mitigate the impact of overconfidence. We investigate the influential factors related to this and discover that, while preserving semantic information, greater data diversity enhances the effectiveness of augmentation. Furthermore, the impact of different augmentation strategies varies across different range of application. Considering parameter transferability and usability, the random combination of augmentations is a promising choice.
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。