Cayley-Math / components /openai-math /validation.json
AvrovaDonz's picture
Add OpenAI mathematics component (part 3)
23610b3 verified
Raw History Blame
1.37 kB
{
"status": "passed",
"source_commit": "4eaee87548250b2682affda7d05d947c05554bae",
"current_papers": 721,
"source_records": 22374,
"source_files_hashed": 8966,
"worker_model": "deepseek-v4.1-flash",
"worker_batches": 74,
"checks": [
"all_delivered_file_hashes_and_sizes",
"721_current_27_previous_3_withdrawn_inventory",
"all_721_real_omp_deepseek_responses_and_input_hashes",
"source_and_macro_provenance_local_files",
"lossless_jsonl_parquet_and_exact_dedup_audit",
"connected_family_group_and_paper_split_separation",
"frozen_worker_observed_flags_and_final_extraction_flags_distinguished",
"separate_ai_metadata_summary_transport",
"all_packed_tokens_reencoded_source_shift_padding_masks_and_attribution"
],
"packed": {
"train": {
"rows": 19145,
"tokens": 39207407,
"loss_bearing_labels": 39185361
},
"dev": {
"rows": 141,
"tokens": 287474,
"loss_bearing_labels": 287311
},
"test": {
"rows": 146,
"tokens": 297055,
"loss_bearing_labels": 296890
}
},
"mathematical_verification": "not_performed",
"limitations": [
"Worker metadata review uses abstracts, headings and bounded excerpts.",
"No theorem correctness assessment or Lean/Comparator compilation.",
"No near-duplicate or Phase 1 overlap assessment."
]
}