Dataset

harbor-eval/swe-issues-clean

1.6k

Cleaned, license-checked GitHub issues with reproducible test harnesses.

codingevalverified

24k
rows: 3.1 GB
size: 31.9k
downloads: 61.0k
eval runs: 1d ago
updated

Dataset card

Cleaned, license-checked GitHub issues with reproducible test harnesses.

Records are deduplicated and license-checked before publishing. Each row carries its source trajectory id, so any example can be traced back to the run, task, and verifier score that produced it.

Format: JSONL / Parquet
Schema: trajectory-v2
Splits: train · eval
Provenance: Verified trials

Use this dataset

CLI

$tune pull harbor-eval/swe-issues-clean

Python

from tune import datasets

ds = datasets.load("harbor-eval/swe-issues-clean", split="train")
for row in ds.take(3):
    print(row["trajectory_id"], row["reward"])