LLM & Models
Our framework for reporting model misalignment
OpenAI Blog
OpenAI shares a framework for tracking, investigating, and disclosing model misalignment, alongside six reports of unexpected or concerning model behavior.
We link to and credit the original publisher. Summary is provided for preview purposes.