Mistral's open model Shieldstral matches much larger safety models at a fraction of the size
AI Summary
Mistral's Shieldstral model, which utilizes natural language yes-or-no questions for safety checks, demonstrates comparable performance to larger safety models while being significantly smaller in size. Users can customize safety criteria in real-time, and the model is capable of running locally.
From the source
Mistral's new 3B Shieldstral model checks AI inputs and outputs for safety violations using natural language yes-or-no questions instead of fixed categories. It matches models seven times its size in some benchmarks. Operators can set their own criteria at runtime rather than rely on a third party's category system, and the model can run locally. The article Mistral's open model Shieldstral matches much larger safety models at a fraction of the size appeared first on The Decoder.
The full text couldn't be loaded here (the source may require a subscription).
View original at The Decoder