Work

Interpreto: An Explainability Library for Transformers

Year: 2025

Type: preprint

Abstract: Interpreto is an open-source Python library for interpreting HuggingFace language models, from early BERT variants to LLMs. It provides two complementary families of methods: attribution methods and concept-based explanations. The library bridges recent research and practical tooling by exposing explanation workflows through a unified API for both classification and text generation. A key differentiator is its end-to-end concept-based pipeline (from activation extraction to concept learning, interpretation, and scoring), which goes beyond feature-level attributions and is uncommon in existing libraries. See GitHub: https://github.com/FOR-sight-ai/interpreto and the demo website: https://for-sight-ai.github.io/interpreto-demo/.

Source: HAL (Le Centre pour la Communication Scientifique Directe)

DOI: 10.48550/arxiv.2512.09730

Authors: Poché, Antonin, Mullor, Thomas, Sarti, Gabriele, Boisnard, Frédéric, Friedrich, Corentin, Claye, Charlotte, Hoofd, François, Bernas, Raphael, Asher, Nicholas, Hudelot, Céline, Jourdan, Fanny

Institutions: Institut de Recherche en Informatique de Toulouse, IRT Saint Exupéry, University of Groningen, CentraleSupélec, Mathématiques et Informatique pour la Complexité et les Systèmes

Language: English

Topic: Explainable Artificial Intelligence (XAI)

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

Funder: Agence Nationale de la Recherche

Awards: G1190173965, Artificial and Natural Intelligence Toulouse Institute, SAINT EXUPERY, G7240651115

OpenAlex ID: W7114909970

Cites:

Cited by:

Related to: