发现权重衰减会不公平地削弱低频词，影响模型对稀有词汇的处理能力。
The Fair Language Model Paradox
- 分析训练时每个词元的表现差异，发现权重衰减导致低频词被系统性贬低
- 在270M至3B参数模型上验证，权重衰减越高，低频词损失越严重
- 提醒开发者关注词元级公平性，尤其对语言中占多数的罕见词
大型语言模型广泛应用于实际场景，但其在词元层面的训练动态仍不明确。评估通常依赖批处理级别的聚合损失，忽略了由（一）词元级动态差异和（二）超参数引入的结构偏差所导致的细微偏差。尽管权重衰减常用于稳定训练，我们实证发现它会在词元层面悄悄引入性能偏差。在从270M到3B参数的不同数据集规模、模型架构和尺寸下，均显示随着权重衰减增加，低频词被不成比例地贬值。这尤为令人担忧，因为大多数语言中，低频词构成了词元分布的绝大部分，亟需新的正则化技术以确保所有词元的公平性。
原文摘要 · Abstract (English)
Large Language Models (LLMs) are widely deployed in real-world applications, yet little is known about their training dynamics at the token level. Evaluation typically relies on aggregated training loss, measured at the batch level, which overlooks subtle per-token biases arising from (i) varying token-level dynamics and (ii) structural biases introduced by hyperparameters. While weight decay is commonly used to stabilize training, we reveal that it silently introduces performance biases detectable only at the token level. In fact, we empirically show across different dataset sizes, model architectures and sizes ranging from 270M to 3B parameters that as weight decay increases, low-frequency tokens are disproportionately depreciated. This is particularly concerning, as these neglected low-frequency tokens represent the vast majority of the token distribution in most languages, calling for novel regularization techniques that ensure fairness across all available tokens.
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。