改进大模型训练数据归属，提升溯源准确性。
Enhancing Training Data Attribution for Large Language Models with Fitting Error Consideration
- 通过消除模型知识偏差和训练拟合噪声，优化影响函数。
- 在多个模型上平均AUC达91.64%，显著优于现有方法。
- 适用于不同来源与规模模型，通用性强且可扩展。
大型语言模型的黑箱特性给结果解释带来挑战，影响数据知识产权保护与幻觉溯源。训练数据归属（TDA）方法被视为有效解决方案。现有TDA方法多依赖影响函数，假设模型达到最小经验风险，但实际训练中该条件难以满足，拟合误差会降低溯源精度。本文提出一种新方法DDA，通过消除模型预训练阶段的知识偏差（去偏策略）和利用平滑技术减少因训练拟合程度差异导致的影响分数波动（去噪策略），从而增强影响函数性能。实验表明，该方法在多个模型（LLaMA2、QWEN2、Mistral）上平均AUC达到91.64%，显著优于现有方法，具备良好的泛化性与可扩展性。
原文摘要 · Abstract (English)
The black-box nature of large language models (LLMs) poses challenges in interpreting results, impacting issues such as data intellectual property protection and hallucination tracing. Training data attribution (TDA) methods are considered effective solutions to address these challenges. Most recent TDA methods rely on influence functions, assuming the model achieves minimized empirical risk. However, achieving this criterion is difficult, and sourcing accuracy can be compromised by fitting errors during model training. In this paper, we introduce a novel TDA method called Debias and Denoise Attribution (DDA), which enhances influence functions by addressing fitting errors. Specifically, the debias strategy seeks to improve the performance of influence functions by eliminating the knowledge bias present in the base model before fine-tuning, while the denoise strategy aims to reduce discrepancies in influence scores arising from varying degrees of fitting during the training process through smoothing techniques. Experimental results demonstrate that our method significantly outperforms existing approaches, achieving an averaged AUC of 91.64%. Moreover, DDA exhibits strong generality and scalability across various sources and different-scale models like LLaMA2, QWEN2, and Mistral.
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。