Gemini 1.5
Google DeepMind — Gemini
rich detail 9 method fields published
Evaluation
Representational harms in audio-to-text (WER across subgroups)
Comparative ASR performance (WER) on AAVE vs SAE longform speech (internal dataset) and male vs female speech (Mozilla Common Voice), compared to USM
How it was run
Open a release to see the setup, scoring, and other details that the lab published. Only fields the lab actually disclosed are shown.