Abstract
Recent studies introduced effective compression techniques for Large Language Models (LLMs) via post-training quantization or low-bit weight representation. Although quantized weights offer storage efficiency and allow for faster inference, existing works have indicated that quantization might compromise performance and exacerbate biases in LLMs. This study investigates the confidence and calibration of quantized models, considering factors such as language model type and scale as contributors to quantization loss. Firstly, we reveal that quantization with GPTQ to 4-bit results in a decrease in confidence regarding true labels, with varying impacts observed among different language models. Secondly, we observe fluctuations in the impact on confidence across different scales. Finally, we propose an explanation for quantization loss based on confidence levels, indicating that quantization disproportionately affects samples where the full model exhibited low confidence levels in the first place.
Abstract (translated)
近年来，通过后训练量化或低位权重表示，有效压缩技术已经引入了大型语言模型（LLMs）。尽管量化权重具有存储效率，能够加速推理，但已有研究表示，量化可能会影响性能，甚至加剧LLMs中的偏见。本研究调查了量化模型的置信度和校准度，考虑了诸如语言模型类型和规模等因素对量化损失的贡献。首先，我们发现，将GPTQ量化到4位会导致关于真实标签的置信度下降，不同语言模型上观察到的影响有所不同。其次，我们观察到不同缩放级别上置信度对的影响波动。最后，我们提出了一个基于置信度的量化损失解释，表明量化 disproportionately影响在训练过程中首次表现出低置信度的样本。
URL
https://arxiv.org/abs/2405.00632
