Interpreto: An Explainability Library for Transformers
A new open-source Python library named Interpreto has been released to aid in the explainability of Transformer-based language models, including BERT and larger LLMs. Developed by Antonin Poché, the library offers both attribution methods and concept-based explanations through a unified API for text classification and generation tasks. A notable feature is its comprehensive concept-based pipeline, which extends beyond typical feature-level attributions. AI
IMPACT Provides researchers and developers with tools to better understand and debug Transformer-based language models.