OpenAI updates its Model Spec around five core values, now calls the model "GPT"
The updated Model Spec is reorganized around faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care, retiring most detailed style guidance.
OpenAI updated its Model Spec, the public description of how it wants its models to behave, reorganizing it around five values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care.
Key changes:
-
A new section on faithfulness to human intent, incorporating an updated chain of command and a new "consider context carefully" subsection.
-
New defaults on supporting human agency and following relevant social norms; expanded guidance on controlling and communicating side effects.
-
Most detailed style guidance retired; stronger defaults such as approaching questions objectively are kept.
-
The Spec now refers to the model as "GPT" rather than "the assistant".
-
OpenAI notes its models do not yet fully reflect the Spec and invites feedback.