Beyond Transcription: Mechanistic Interpretability in ASR

Neta Glazer, Yael Segal-Feldman, Hilit Segev +6

Interpretability methods have recently gained significant attention, particularly in the context of large language models, enabling insights into linguistic representations, error…