nano-gpt-data-curation
gpt-5.6-sol
nano-gpt-data-curation — graded by the task's verifier against a measured baseline.
- backend
- modal
- variant
- codex azure_ai/gpt-5.6-sol
- metric
- held-out multi-domain perplexity
- actions
- 217
- duration
- 1h 4m
- recorded
- 2026-08-03
run iddclm-nano-curation__5zjKKAd
Score
reward
355.460
Summary
Completed all deliverables:
- selection.json — 182,014 unique valid IDs; verified exact 12M-token packing
- curate.py — fully reproducible curation
- claim.md — hypothesis, mechanism, falsification, and transfer
Frozen-script dev perplexity: 350.79, versus 465.53 for random selection.