DeepScaleR-Easy-10k / README.md
JWei05's picture
README: x16 replication is a local prep step, not a hub file
0c3e81d verified
|
Raw
History Blame Contribute Delete
1.32 kB
metadata
license: mit

DeepScaleR-Easy-10k

RL training set of 10,000 DeepScaleR questions that Gemma-3-4B-IT solved 4/4 (temp 1.0, 4 samples, LENIENT math_verify grading), split train 9,500 / val 500.

Creation (seed 42 throughout):

  1. Gemma-3-4B-IT generated 4 responses per DeepScaleR question (40,315 rows).
  2. Dedup v2: questions grouped under max-strength normalization (case, whitespace, prose punctuation, LaTeX formatting); duplicate groups whose gold answers are not all math_verify-equivalent were removed entirely (167 groups / 567 rows — label conflicts); one random copy kept per surviving group -> 38,796 unique questions. Script: rl-distill-scripts/data/dedup_deepscaler_it_gen.py.
  3. 10,000 sampled uniformly from the lenient 4/4 bucket (10,053 available).
  4. 500 sampled uniformly as validation; remaining 9,500 are train.

Files: *_train.parquet and *_val500.parquet (500 unique held-out questions). For mean@16 validation, replicate the val rows 16x locally (verl samples one response per row). verl format: data_source "math" (routes to the repo's math_verify scorer), prompt = single user message, reward_model.ground_truth, extra_info carries lenient/strict pass counts.

Built by rl-distill-scripts/data/build_deepscaler_easy_medium.py in JasonWei05/rl-distill.