Focus-LIME: Surgical Interpretation of Long-Context Large Language Models via Proxy-Based Neighborhood Selection
This addresses the need for precise interpretation in high-stakes tasks like legal auditing and code debugging, representing an incremental improvement over existing local model-agnostic methods.
The paper tackles the problem of achieving surgical feature-level interpretation for long-context large language models, which is hindered by attribution dilution in existing methods, and proposes Focus-LIME, a framework that restores tractability and provides faithful explanations as demonstrated in empirical evaluations.
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.