Open Models Almanac
Methodspost-trainingreinforcement learning algorithm

single mention · post-training

SAO with compaction

A named RL strategy carried over from GLM-5.2; the evidence does not specify its mechanism.

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 SAO with compaction (RL strategy).

Filed under post-training › reinforcement learning algorithm