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