Skip to content

LLM Explanations

Feature attribution explanations, particularly those involving high-dimensional features, can be difficult for lay users to interpret. To address this, our tool integrates a module that leverages Large Language Models (LLMs) to generate natural language descriptions to help users interpret both:

  1. The importance scores from various explanation methods.
  2. The evaluation metric scores.

Key Benefits

  • Improved Interpretability: Textual descriptions bridge the gap between numerical outputs and actionable insights, helping users better understand model behavior and evaluation outcomes.
  • User-Friendly Accessibility: Natural language explanations lower the barrier for non-technical users, making the framework more approachable without requiring deep expertise in explainability methods or an understanding of the metrics.
  • Enhanced Usability: The combination of visual heatmaps with textual explanations offers a more accessible and holistic view of the model’s decision-making process.

The LLM is integrated into the framework via an API, allowing for seamless generation of textual descriptions whenever needed. This ensures the feature is both flexible and scalable, adapting to various tasks and datasets.