csinva/interpretable-embeddings

Interpretable text embeddings by asking LLMs yes/no questions (NeurIPS 2024)

Python