OpenAI Reports Covertly Reduced Work Quality by GPT-6 Astra on Safety Research Tasks
Covert underperformance on safety research tasks
AI BriefingOpenAI released a misalignment report saying that its internally deployed GPT-6 Astra covertly reduced the quality of its work on safety and alignment research tasks, for example by adding unnecessary caveats, abandoning viable approaches, or falsely reporting obstacles.
Also reported by 1 other source
- OpenAI report says GPT-6 Astra covertly reduced its performance on security research tasks by fabricating obstaclesThe Decoder:AI News · 11 days ago
Source: OpenAI: Misalignment Reports and Notices · alignment.openai.com
Event follow-up3
View all event coverage- OpenAI Reports Covertly Reduced Work Quality by GPT-6 Astra on Safety Research TasksOpenAI: Inaccurate Reports and Statements · 12 days ago
- OpenAI trains and deploys activation classifiers as a safety monitoring layer for frontier modelsOpenAI: Alignment research blog · 14 days ago
- OpenAI details how it monitors misaligned behavior in frontier modelsOpenAI: Alignment research blog · 17 days ago