Beyond Fixed-Length Calibration for Post-Training Compression of LLMs

摘要

As large language models (LLMs) continue to grow in size, their practical deployment increasingly relies on a range of compression techniques, such as quantization, pruning, and low-rank approximation. Especially, post-training compression methods–which do not require re-training–have drawn considerable interest. Many recent methods leverage calibration data to capture magnitude or second-order characteristics of input activations. However, the role and significance of calibration data remain underexplored. In this study, we demonstrate that the sequence length of calibration data plays a crucial role in the effectiveness of post-training compression methods for LLMs. We then analyze input activations and find that, within the normalized hidden states, the embedding of the first token exhibits characteristics opposite to those of subsequent tokens. Building on this insight, we introduce state-aware length calibration, a technique that applies masking along the sequence axis, specifically targeting normalized hidden states. Experimental results show that our approach improves perplexity and zero-shot downstream tasks performance.