Open Models Almanac
Methodsdata curationdata mixture & curriculum

single mention · data curation

ProtocolQA-aligned RL training datasets

Using RL datasets aligned with ProtocolQA; the evidence does not specify the alignment procedure further.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See data mixture & curriculum for the methods filed beside it.

Also called Aligned RL datasets with ProtocolQA.

Filed under data curation › data mixture & curriculum