提出首个关注公平性的量化方法，减少大模型生成偏见内容的风险。
Fair-GPTQ: Bias-Aware Quantization for Large Language Models
- 在量化目标中加入群体公平性约束，引导舍入过程降低偏见。
- 4比特量化下保留90%以上基线准确率，公平性显著优于半精度模型。
- 适合关注模型伦理、需部署低资源大模型的开发者使用。
生成式语言模型的高内存需求促使人们关注量化技术，该技术通过将模型权重映射到低精度整数来降低内存占用。然而，近期实证研究表明，尽管高效，量化可能增加生成偏见输出的概率，并降低公平性基准上的表现。本文首次建立量化与模型公平性之间的联系，通过在量化目标中引入显式的群体公平性约束，提出Fair-GPTQ——首个专为减少大语言模型不公平性而设计的量化方法。该约束引导舍入操作向保护群体更少偏见的文本生成方向学习。具体聚焦于职业刻板印象及涉及性别、种族和宗教的歧视性语言。Fair-GPTQ对性能影响极小，在零样本基准上保持至少90%的基线准确率，相较于半精度模型显著降低不公平性，同时保留4比特量化带来的内存与速度优势。
原文摘要 · Abstract (English)
The high memory demands of generative language models have drawn attention to quantization, which reduces memory usage by mapping model weights to lower-precision integers. However, recent empirical studies show that, while efficient, quantization can increase the likelihood of generating biased outputs and degrade performance on fairness benchmarks. In this work, we draw new links between quantization and model fairness by adding explicit group-fairness constraints to the quantization objective and introduce Fair-GPTQ, the first quantization method explicitly designed to reduce unfairness in large language models. The added constraints guide the learning of the rounding operation toward less-biased text generation for protected groups. Specifically, we focus on stereotype generation involving occupational bias and discriminatory language spanning gender, race, and religion. Fair-GPTQ has minimal impact on performance, preserving at least 90% of baseline accuracy on zero-shot benchmarks, reduces unfairness relative to a half-precision model, and retains the memory and speed benefits of 4-bit quantization.
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。