Open Models Almanac
Methodsinference & servingagentic scaffolding

specific method · filed under inference & serving

Role-based instruction hierarchy

Instruction conflicts are resolved according to the priority order system, developer, user, assistant, then tool.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See agentic scaffolding for the methods filed beside it.

Also called role hierarchy.

sources
2
model
1
lab adopt it
1
strongest
core

Documented in

Evidence

2 quotes from 2 sources, one card per model, each opening with its strongest.

How sources treat it
used 1core 1

gpt-oss-120b

OpenAIadopted2 sources

These roles also represent the information hierarchy that the model applies in case there are any instruction conflicts: system > developer > user > assistant > tool

coreinference servingin gpt-ossOpenAI
1 more from this model

Using these roles, the models follow a role-based information hierarchy to resolve instruction conflicts: System > Developer > User > Assistant > Tool.

usedpost trainingin gpt-oss-120b & gpt-oss-20bOpenAI
↑