RCStat: A Statistical Framework for using Relative Contextualization in Transformers

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Mahapatra, Debabrata, Agarwal, Shubham, Saxena, Apoorv, Mitra, Subrata
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916809669607424
author Mahapatra, Debabrata
Agarwal, Shubham
Saxena, Apoorv
Mitra, Subrata
author_facet Mahapatra, Debabrata
Agarwal, Shubham
Saxena, Apoorv
Mitra, Subrata
contents Prior work on input-token importance in auto-regressive transformers has relied on Softmax-normalized attention weights, which obscure the richer structure of pre-Softmax query-key logits. We introduce RCStat, a statistical framework that harnesses raw attention logits via Relative Contextualization (RC), a random variable measuring contextual alignment between token segments, and derive an efficient upper bound for RC. We demonstrate two applications: (i) Key-Value compression, where RC-based thresholds drive adaptive key-value eviction for substantial cache reduction with minimal quality loss; and (ii) Attribution, where RC yields higher-fidelity token-, sentence-, and chunk-level explanations than post-Softmax methods. Across question answering, summarization, and attribution benchmarks, RCStat achieves significant empirical gains, delivering state-of-the-art compression and attribution performance without any model retraining.
format Preprint
id arxiv_https___arxiv_org_abs_2506_19549
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle RCStat: A Statistical Framework for using Relative Contextualization in Transformers
Mahapatra, Debabrata
Agarwal, Shubham
Saxena, Apoorv
Mitra, Subrata
Computation and Language
Artificial Intelligence
Machine Learning
Prior work on input-token importance in auto-regressive transformers has relied on Softmax-normalized attention weights, which obscure the richer structure of pre-Softmax query-key logits. We introduce RCStat, a statistical framework that harnesses raw attention logits via Relative Contextualization (RC), a random variable measuring contextual alignment between token segments, and derive an efficient upper bound for RC. We demonstrate two applications: (i) Key-Value compression, where RC-based thresholds drive adaptive key-value eviction for substantial cache reduction with minimal quality loss; and (ii) Attribution, where RC yields higher-fidelity token-, sentence-, and chunk-level explanations than post-Softmax methods. Across question answering, summarization, and attribution benchmarks, RCStat achieves significant empirical gains, delivering state-of-the-art compression and attribution performance without any model retraining.
title RCStat: A Statistical Framework for using Relative Contextualization in Transformers
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2506.19549