Skip to content
nano-gpt-data-curation

claude-opus-4-8

nano-gpt-data-curation — graded by the task's verifier against a measured baseline.

backend
modal
variant
claude claude-opus-4-8
metric
held-out multi-domain perplexity
actions
49
duration
40m
recorded
2026-08-04

run iddclm-nano-curation__YJRWvfs

Score

reward
368.690
compare against every run on this task →

Summary

All deliverables are in place. Summary of what I did and found:

Approach

The disclosed dev target (multi_dev.npy) decodes to a broad high-quality English mix (Wikipedia, web prose, news, technical Q&A). I used it directly as the positive class of a target-vs-pool classifier:

  • Positives: the dev target text decoded back to raw text (normalizing WikiText @-@/@,@ artifacts).
  • Negatives: a random pool sample.
  • Features: hashed word uni+bigrams, length-normalized; L2 logistic regression trained on the H100 (pure torch — no sklearn/scipy available, no internet).
  • Prefilter: Gopher/C4-style junk rules (length, alpha ratio, mean word length, bullet/line-repetition spam).

Key empirical finding (the interesting part)

Ranking the selection by classifier score (best-first) was worse than random (500 vs 465 dev PPL) — the first 12M tokens became one homogeneous register. The winning move was to filter, then shuffle. Sweeping the keep-fraction produced a clear U-shape:

selectiondev PPL
random baseline465
rank best-first (narrow)500
keep 30%, shuffled443
keep 55%423
keep 70%~412
keep 85%, shuffled~396 (365–424 across shuffles)
keep 100% (junk-only)432

Final selection: drop the off-target bottom 15%, shuffle the rest — 142,965 valid unique ids, fills the 12M budget, ~15% below the random baseline on dev.

Deliverables

  • /workspace/submission/selection.json — 142,965 ordered ids (valid, unique, in-range, fills budget).
  • /workspace/submission/curate.py — reproducible, stated-criterion script.
  • /workspace/claim.md — hypothesis (diversity beats purity), mechanism (the non-monotonic U-shape as an observable prediction), falsification tests, and transfer conditions.

One honest caveat: the shuffle induces ±20 PPL variance, so the single best 365 run was partly luck; the submitted ordering is a representative keep-85% draw with expected dev PPL ~396.

Transcript