Do Emergent Abilities Exist in Quantized Large Language Models: An Empirical Study

TL;DR

This study investigates how quantization affects the emergent abilities of large language models, finding that 4-bit models retain these abilities while 2-bit models suffer significant performance loss, with potential for further optimization.

Contribution

It provides the first detailed empirical analysis of how different levels of quantization impact emergent capabilities in LLMs, including insights for improving low-bit models.

Findings

4-bit quantized models retain emergent abilities

2-bit quantized models show severe performance degradation

Fine-tuning can partially recover performance in low-bit models

Abstract

Despite the superior performance, Large Language Models~(LLMs) require significant computational resources for deployment and use. To overcome this issue, quantization methods have been widely applied to reduce the memory footprint of LLMs as well as increasing the inference rate. However, a major challenge is that low-bit quantization methods often lead to performance degradation. It is important to understand how quantization impacts the capacity of LLMs. Different from previous studies focused on overall performance, this work aims to investigate the impact of quantization on \emph{emergent abilities}, which are important characteristics that distinguish LLMs from small language models. Specially, we examine the abilities of in-context learning, chain-of-thought reasoning, and instruction-following in quantized LLMs. Our empirical experiments show that these emergent abilities still exist in 4-bit quantization models, while 2-bit models encounter severe performance degradation on the test of these abilities. To improve the performance of low-bit models, we conduct two special experiments: (1) fine-gained impact analysis that studies which components (or substructures) are more sensitive to quantization, and (2) performance compensation through model fine-tuning. Our work derives a series of important findings to understand the impact of quantization on emergent abilities, and sheds lights on the possibilities of extremely low-bit quantization for LLMs.