Open Models Almanac
Methodspost-trainingsupervised fine-tuning

single mention · post-training

Mask tool responses from the SFT loss

Limits agent traces in the training blend and excludes tool responses from the loss so training targets the model's own turns.

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