量化会微妙影响大模型偏见，压缩越强越易加剧刻板印象。
How Quantization Shapes Bias in Large Language Models
- 对比权重与激活量化，分析其对多种偏见的影响机制。
- 量化虽降毒性，但加剧刻板印象与不公平，尤其在强压缩下。
- 结果跨模型架构与人群一致，提示部署时需权衡效率与伦理。
本研究系统评估了量化对大语言模型偏见的影响，重点关注其对不同人口统计子群体的作用。聚焦权重和激活量化策略，考察其在涵盖刻板印象、公平性、毒性及情感等多类偏见上的表现。基于13个基准测试，采用概率与生成文本双类指标，评估不同架构家族与推理能力的模型。结果表明：量化对偏见影响复杂——虽可降低模型毒性且对情感无显著影响，但在生成任务中往往轻微增加刻板印象与不公平性，尤其在激进压缩条件下。该趋势在多数人口类别与模型类型中保持一致，但程度随具体设置变化。总体说明，在实际应用中需审慎平衡量化效率与伦理考量。
原文摘要 · Abstract (English)
This work presents a comprehensive evaluation of how quantization affects model bias, with particular attention to its impact on individual demographic subgroups. We focus on weight and activation quantization strategies and examine their effects across a broad range of bias types, including stereotypes, fairness, toxicity, and sentiment. We employ both probability- and generated text-based metrics across 13 benchmarks and evaluate models that differ in architecture family and reasoning ability. Our findings show that quantization has a nuanced impact on bias: while it can reduce model toxicity and does not significantly impact sentiment, it tends to slightly increase stereotypes and unfairness in generative tasks, especially under aggressive compression. These trends are generally consistent across demographic categories and subgroups, and model types, although their magnitude depends on the specific setting. Overall, our results highlight the importance of carefully balancing efficiency and ethical considerations when applying quantization in practice.
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。