Ready-to-train data
Graded rollouts by agent. Rows, size and what the analyzer measured are on the card.
pip install zeroproof·rows = zps.pull("ds_…")no key for public sets
RL rollouts from the quant coding agent, 2026-09-11 run.
zps.pull("ds_ba9f48fba706f0e3")
Production support-agent traces, graded. Analyzer regime TRAIN.
zps.pull("ds_550a2b3ba46aac2b")
publish yours: data.push(name, agent="…", publish=True)·docs