Open Models Almanac
Methodspost-trainingreward modelling

single mention · post-training

Per-token tool-error reward shaping

Assigns shaping at the failing tool-use token to focus credit assignment and discourage malformed calls.

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