Include 100-200 row samples so the evaluation scripts can be run without regenerating the full datasets. |
||
|---|---|---|
| .. | ||
| __init__.py | ||
| build_real_ffs_dataset.py | ||
| dataset.md | ||
| lm_dataset.py | ||
| parse_all_r20_docs.py | ||
| patent_breakdown_200.jsonl | ||
| ran2_benchmark_test_50.jsonl | ||
| ran2_expert_sft_100.jsonl | ||
| test_ffs_blind.jsonl | ||