Datasetpublic

harman/tts

TTS synthetic programming-question dataset This public dataset contains the frozen set of 2,865 accepted programming questions and their materialized verifier tests. Accepted question bundle accepted_bundle/accepted-questions-2865.tar.zst contains all 42,975 accepted question files: statements, package JSON, three public examples, generators, validators, reference solutions, brute-force solutions, verification records, provenance records, and GPT-OSS hardness… See the full description on the dataset page: https://huggingface.co/datasets/harman/tts.

sourceHugging Faceotherupdated 18d agoView on Hugging Face

0likes764downloads

Dataset Card

TTS synthetic programming-question dataset

This public dataset contains the frozen set of 2,865 accepted programming questions and their materialized verifier tests.

Accepted question bundle

accepted_bundle/accepted-questions-2865.tar.zst contains all 42,975 accepted question files: statements, package JSON, three public examples, generators, validators, reference solutions, brute-force solutions, verification records, provenance records, and GPT-OSS hardness sidecars.

Hidden test fixtures

full_tests/fixture50/shards/ contains 39 compressed shards. Every question has exactly 50 hidden input/output pairs:

—30 small tests, seeds 0..29
—20 large tests, seeds 100000..100019

Across all questions this is 143,250 hidden test pairs. These are the pre-existing materialized verification fixtures. No additional hidden tests beyond these 50 per question are included.

Each shard contains paths of the form:

text

fixture50/<model_tag>/<generation_id>/000.in
fixture50/<model_tag>/<generation_id>/000.out
fixture50/<model_tag>/<generation_id>/manifest.json

Extract Zstandard tar archives with:

bash

tar --zstd -xf <archive>.tar.zst

See full_tests/fixture50/manifests/fixture50_plan.json for question-to-shard mapping, fixture50_state.json for per-shard SHA-256 values, and manifests/final_hf_audit.json for the final remote verification report.