CLFeb 3, 2025

Massive Values in Self-Attention Modules are the Key to Contextual Knowledge Understanding

arXiv:2502.01563v444 citationsh-index: 17Has CodeICML
Originality Incremental advance
AI Analysis

This provides insights into LLM mechanisms for researchers and practitioners, though it is incremental as it builds on existing understanding of attention and RoPE.

The paper identifies that concentrated massive values in attention queries and keys, but not values, are critical for contextual knowledge understanding in transformer-based LLMs, and shows that ignoring them causes performance drops in tasks requiring contextual understanding.

Large language models (LLMs) have achieved remarkable success in contextual knowledge understanding. In this paper, we show that these concentrated massive values consistently emerge in specific regions of attention queries (Q) and keys (K) while not having such patterns in values (V) in various modern transformer-based LLMs (Q, K, and V mean the representations output by the query, key, and value layers respectively). Through extensive experiments, we further demonstrate that these massive values play a critical role in interpreting contextual knowledge (knowledge obtained from the current context window) rather than in retrieving parametric knowledge stored within the model's parameters. Our further investigation of quantization strategies reveals that ignoring these massive values leads to a pronounced drop in performance on tasks requiring rich contextual understanding, aligning with our analysis. Finally, we trace the emergence of concentrated massive values and find that such concentration is caused by Rotary Positional Encoding (RoPE), which has appeared since the first layers. These findings shed new light on how Q and K operate in LLMs and offer practical insights for model design and optimization. The Code is Available at https://github.com/MingyuJ666/Rope_with_LLM.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes