OpenAI rewrites its Model Spec around five values, adding guidance on context, oversight and agency
Synopsis
OpenAI has updated its Model Spec, the public description of how it wants its models to behave, reorganizing the guidance around five values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care. The previous version spread these commitments across separate principles alongside extensive detailed style guidance; the revision adds guidance on context, oversight and agency, adjusts some behavioral defaults, and retires most of the style detail, while calling the model "GPT" rather than "the assistant." OpenAI says its models do not yet fully reflect the Spec and that results from related training work will come in future updates.
Interpretation
OpenAI has updated the Model Spec, reorganizing its behavioral guidance around five values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care. Previously these commitments were spread across separate entries mixed with extensive detailed style guidance; the revision brings them together to explain why the Spec asks models to behave as it does. OpenAI states that the revision adds guidance on context, oversight and agency, adjusts some behavioral defaults, and retires most of the detailed style guidance.
The new Spec asks models to go beyond following instructions literally, to understand what OpenAI, developers and users actually intend, and to act wisely when that understanding is incomplete. The Spec adds a section on faithfulness to human intent and incorporates an updated chain of command explaining which instructions apply and how conflicts are resolved. A new subsection on considering context carefully explains how conversations, memories and other contextual evidence should inform the model's understanding of people's intentions and preferences; the Spec also adds defaults on supporting human agency and following relevant social norms, and expands guidance on controlling and communicating side effects.
The Spec now refers to the model as "GPT" rather than "the assistant," reflecting the broadening range of use cases and deployment settings. OpenAI describes the update as a first step in making the values and reasoning behind the Spec more explicit. OpenAI states that the Model Spec describes the behavior it is working toward, that its models do not yet fully reflect it, and that it will share results and lessons from related training work in future updates.
Perspective
The revision concerns the text of OpenAI's Model Spec, its public description of how models should behave, and addresses readers who use, build or evaluate such models; OpenAI says it will keep improving both the Spec and the models based on research, real-world use and feedback.
OpenAI states explicitly that the Model Spec describes the behavior it is working toward and that its models do not yet fully reflect it; results and lessons from the related training methods will be shared in future updates.