helloelwin/Qwen3-32B_10episodes_comparisons_full

Qwen3-32B_10episodes_comparisons_full This is a pairwise comparison dataset created from SWE-bench evaluation results. Files Qwen3-32B_10episodes_comparisons_full_comparison_pairs.jsonl: JSONL file containing comparison pairs Metadata { "dataset_name": "Qwen3-32B_10episodes_comparisons_full", "model_name": "Qwen3-32B", "num_episodes": 10, "episodes_used": [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 ]… See the full description on the dataset page: https://huggingface.co/datasets/helloelwin/Qwen3-32B_10episodes_comparisons_full.