Gemini 1.0
Google DeepMind — Gemini
rich detail 10 method fields published
Evaluation
Image-to-text caption quality representational harms (CIDEr across groups)
Whether images of people are described with similar quality (CIDEr) for different gender appearances and skin tones, following Zhao et al. (2021)
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.