Open Models Almanac
Methodspost-trainingreinforcement learning algorithm

single mention · post-training

CISPO-style REINFORCE objective

A REINFORCE loss using CISPO-style importance-sampling-ratio clipping.

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 CISPO-style REINFORCE loss.

Filed under post-training › reinforcement learning algorithm