Search
Have an existing account?
Sign In
Tag: Interpretability
An AI edited a small model’s weights and doubled its math score
MIT researchers say GPT-6 Astra changed 0.016 percent of Qwen's parameters, taking MATH-500 from 7.9 percent to 48.1…
Goodfire watches agents from the inside to cut monitoring costs
Interpretability startup Goodfire has launched monitors that read a model's internal signals instead of paying a second AI…
Baseten’s Base Labs recruits Hugging Face and Goodfire for open model safety
Baseten has launched a safety infrastructure effort for open-weight models with Hugging Face and Goodfire AI.