Open Models Almanac
Methodspost-trainingreinforcement learning algorithm

single mention · post-training

Mask re-rendered context during training

Masks re-rendered context so each generated token is trained on only once.

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

Also called masks this context for the training steps.

Filed under post-training › reinforcement learning algorithm