Open Models Almanac
Methodspost-trainingagentic post-training

single mention · post-training

Multi-harness trajectory training with native behavior preservation

Collecting training trajectories while retaining each harness’s native features, such as subagent spawning, context compaction, and planning scaffolds.

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

Filed under post-training › agentic post-training