Open Models Almanac
Methodspost-trainingreward modelling

single mention · post-training

Rule-based or model-judged trajectory error detection

Uses handcrafted logic or a model-based judge to assess a segment, context, or sequence.

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.

Also called Rule.

Filed under post-training › reward modelling