Work

Interpreto: An Explainability Library for Transformers

Year: 2026

Type: other

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.

Source: Underline Science Inc.

DOI: 10.48448/sbtq-et89

Authors: Association for Computational Linguistics 2026, Asher, Nicholas, Boisnard, Frédéric, Claye, Charlotte, Friedrich, Corentin, Hoofd, François, Hudelot, Céline, Jourdan, Fanny, Mullor, Thomas, Poché, Antonin, Sarti, Gabriele

Institutions: Centre National de la Recherche Scientifique, Université Paris-Saclay, IRT Saint Exupéry

Language: English

Topic: Explainable Artificial Intelligence (XAI)

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7168574792

Cites:

Cited by:

Related to: 10