提出无需反向传播的高效数据估值方法，可批量处理大模型微调数据。
For-Value: Efficient Forward-Only Data Valuation for finetuning LLMs and VLMs
- 仅用一次前向计算，通过隐藏层表示与预测误差对齐度评估数据价值。
- 在检测关键数据和错误标签上效果媲美甚至超越梯度法，速度提升显著。
- 适合需要快速筛选高质量训练数据的大模型开发者使用。
数据估值对提升大语言模型（LLMs）和视觉语言模型（VLMs）的透明性与可问责性至关重要。然而，现有方法多依赖梯度计算，对十亿参数级模型而言计算成本过高，且无法实现批处理并行。本文提出 For-Value，一种仅需前向传播的数据估值框架，可在保持有效性的同时实现高效的批量可扩展估值。基于预训练模型的表达能力，我们理论证明：数据价值可由最后一层的隐藏表示与预测误差之间的对齐程度捕捉。据此，For-Value 采用单一前向传播的闭式公式计算数据价值，完全避免高成本的反向传播，支持大规模批处理。大量实验表明，For-Value 在识别重要数据和误标数据方面表现优于或等同于基于梯度的基线方法，同时实现显著的效率提升。
原文摘要 · Abstract (English)
Data valuation is essential for enhancing the transparency and accountability of large language models (LLMs) and vision-language models (VLMs). However, existing methods typically rely on gradient computations, making them computationally prohibitive for billion-parameter models and precluding batch parallelization. In this work, we introduce For-Value, a forward-only data valuation framework that enables efficient batch-scalable value estimation while maintaining effectiveness. Leveraging the expressive power of pretrained LLMs/VLMs, we theoretically demonstrate that data valuation can be captured by the alignment between the final hidden representations and prediction errors at the last layer. In light of this insight, For-Value computes data value using a simple closed-form expression with a single forward pass, eliminating the need for costly backpropagation and enabling efficient batch calculating at scale. Extensive experiments show that For-Value matches or outperforms gradient-based baselines in detecting influential data and mislabeled data, while achieving significant efficiency improvements.
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。