OpenAI says the incident shows labs need standards for reporting misalignment events, not only model properties described in research papers and system cards.
Need help?
Contact usOpenAI says the incident shows labs need standards for reporting misalignment events, not only model properties described in research papers and system cards.