OpenAI revises its model behavior guidelines

OpenAI reorganized its public Model Spec around five values, with faithfulness to human intent as the foundation. The company said current models do not yet fully reflect the revised framework and that it is exploring training methods to help them apply its principles under human oversight.

• OpenAI reorganized its Model Spec around five values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care.· 1 source
• The company said models should interpret what users, developers, and OpenAI intend and value, rather than simply follow instructions literally; it also said the models should remain subject to human oversight.· 1 source
• OpenAI said current models do not yet fully reflect the updated specification and that it is exploring training methods to help them internalize its principles.· 1 source