Open Models Almanac
Methodspost-trainingsupervised fine-tuning

single mention · post-training

Instruction hierarchy training

Supervises the model to prioritize system instructions over developer instructions and developer instructions over user instructions.

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

Filed under post-training › supervised fine-tuning