Focus-LIME: Surgical Interpretation of Long-Context Large Language Models via Proxy-Based Neighborhood Selection

TL;DR

Focus-LIME introduces a proxy-based, coarse-to-fine framework that enables precise, faithful interpretation of large language models with extensive context windows, addressing attribution challenges in high-stakes tasks.

Contribution

It presents a novel proxy-guided neighborhood selection method that enhances the interpretability of long-context LLMs by enabling surgical feature attribution.

Findings

Effective in long-context benchmarks

Provides faithful and precise explanations

Addresses attribution dilution in high-dimensional features

Abstract

As Large Language Models (LLMs) scale to handle massive context windows, achieving surgical feature-level interpretation is essential for high-stakes tasks like legal auditing and code debugging. However, existing local model-agnostic explanation methods face a critical dilemma in these scenarios: feature-based methods suffer from attribution dilution due to high feature dimensionality, thus failing to provide faithful explanations. In this paper, we propose Focus-LIME, a coarse-to-fine framework designed to restore the tractability of surgical interpretation. Focus-LIME utilizes a proxy model to curate the perturbation neighborhood, allowing the target model to perform fine-grained attribution exclusively within the optimized context. Empirical evaluations on long-context benchmarks demonstrate that our method makes surgical explanations practicable and provides faithful explanations to users.