$ python3 -I tools/experiments/charactereval_release_checks.py <repo> <eprint> <bundle>/results/release_checks.json
# started 2026-10-29T12:00:29+00:00  ended 2026-10-29T12:00:30+00:00  exit 0

R1 formats: {'data/test_data.jsonl': 'list', 'data/id2metric.jsonl': 'dict', 'data/character_profiles.json': 'dict', 'rm_train_data.json': 'list', 'results/generation_trans_chatglm3_6b.jsonl': 'list', 'results/generation_trans_baichuan_7b.jsonl': 'list', 'results/generation_trans_xverse_7b.jsonl': 'list', 'results/generation_trans_internlm_7b.jsonl': 'list', 'results/generation_trans_qwen_7b.jsonl': 'list'}
R2 rm_train_data: 7228 rows, 4241 ids, 4241 of them test ids, 7228 rows with a test example's role and context; (id, metric) pairs 7228 of id2metric's 8032, subset: True; labels {'1': 1361, '2': 1296, '3': 1874, '4': 1783, '5': 914}
R3 frozen outputs: {'chatglm3_6b': {'rows': 8032, 'ids': 4531, 'none_or_error': 0, 'empty': 0}, 'baichuan_7b': {'rows': 8032, 'ids': 4531, 'none_or_error': 0, 'empty': 0}, 'xverse_7b': {'rows': 8032, 'ids': 4531, 'none_or_error': 0, 'empty': 0}, 'internlm_7b': {'rows': 8032, 'ids': 4531, 'none_or_error': 0, 'empty': 0}, 'qwen_7b': {'rows': 8032, 'ids': 4531, 'none_or_error': 0, 'empty': 0}}
R4 example-weighted dimension averages vs printed: max |diff| 0.0628