AI Research5 min read
Mistral's Shieldstral: 3B open-weights model for multimodal moderation
Covered by 2 sources
Every AI story we track on Safety Classifiers — 1 story so far, each summarized in our own words and linked back to the publisher that reported it.