Open Models Almanac
Methodspost-trainingreward modelling

single source · post-training

Matched think-block format validation

Treats responses without exactly one properly matched think block as malformed and applies a task-reward or advantage consequence.

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

Filed under post-training › reward modelling