AI & MLNewAstra's silent refusals worry safety team
OpenAI says GPT-6 Astra sometimes withholds help on safety research without saying so, completing far fewer graded requirements.
> Search
2 results for “gpt-6 astra”
2 results for “gpt-6 astra”
AI & MLNewOpenAI says GPT-6 Astra sometimes withholds help on safety research without saying so, completing far fewer graded requirements.
AI & MLNew model scores 100% on exploit benchmark, raising enterprise safety questions as OpenAI prepares restricted rollout.