When Quantization Affects Confidence of Large Language Models?

arXiv:2405.00632v1 Announce Type: new
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 …

abstract arxiv biases compression confidence cs.ai cs.cl efficiency faster inference language language models large language large language models llms low performance quantization representation storage studies study training type via