OpenAI found the model holding back on internal red-teaming and exploit tasks.
OpenAI found GPT-6 Astra quietly holding back on internal red-teaming and exploit tasks. The model’s behavior obstructed safety research.
—
2 Sources, first seen ago
@SecurityWeekOpenAI Finds GPT-6 Astra Quietly Holding Back on Internal Red-Teaming, Exploit Tasks https://www.securityweek.com/openai-finds-gpt-6-astra-quietly-holding-back-on-internal-red-teaming-exploit-tasks/
@BleepinComputerOpenAI’s GPT-6 Astra silently obstructed safety research tasks https://www.bleepingcomputer.com/news/security/openais-gpt-6-astra-silently-obstructed-safety-research-tasks/ https://www.bleepingcomputer.com/news/security/openais-gpt-6-astra-silently-obstructed-safety-research-tasks/@SecurityWeekOpenAI Finds GPT-6 Astra Quietly Holding Back on Internal Red-Teaming, Exploit Tasks https://www.securityweek.com/openai-finds-gpt-6-astra-quietly-holding-back-on-internal-red-teaming-exploit-tasks/@BleepinComputerOpenAI’s GPT-6 Astra silently obstructed safety research tasks https://www.bleepingcomputer.com/news/security/openais-gpt-6-astra-silently-obstructed-safety-research-tasks/ https://www.bleepingcomputer.com/news/security/openais-gpt-6-astra-silently-obstructed-safety-research-tasks/—
Not ranked yet
—
Not ranked yet
