{ "stages": { "stable:queen-v4-r1-stable": { "wall_seconds": 11067.888490438461, "invocations": 1 }, "branch:queen-v4-r1-x1": { "wall_seconds": 1519.7983586788177, "invocations": 1 }, "branch:queen-v4-r1-x1.5": { "wall_seconds": 2177.17232131958, "invocations": 1 }, "branch:queen-v4-r1-x2": { "wall_seconds": 2837.695327758789, "invocations": 1 }, "sft": { "wall_seconds": 4540.518196582794, "invocations": 1 } }, "gpu": { "count": 1 }, "wall_hours": 6.151, "gpu_hours": 6.151, "usd_per_gpu_hour": 0.34, "cost_usd": 2.09, "runs": { "queen-v4-r1-stable": { "last_step": 3846, "last_val_bpb": 0.9321817332330166, "best_step": 3072, "best_val_bpb": 0.9221891606398556, "sft_from_step": 3072 }, "queen-v4-r1-x1": { "last_step": 2404, "last_val_bpb": 1.2690717874147677, "best_step": 1980, "best_val_bpb": 1.004812328168178, "sft_from_step": 1980 }, "queen-v4-r1-x1.5": { "last_step": 3605, "last_val_bpb": 1.4380317966698843, "best_step": 2970, "best_val_bpb": 0.9712996006066769, "sft_from_step": 2970 }, "queen-v4-r1-x2": { "last_step": 4808, "last_val_bpb": 1.6860155733525417, "best_step": 3960, "best_val_bpb": 0.9732352409739348, "sft_from_step": 3960 } }, "base": { "last_step": 3605, "last_val_bpb": 1.4380317966698843, "best_step": 3605, "best_val_bpb": 1.4380317966698843, "sft_from_step": 3605, "promoted_from": "queen-v4-r1-x1.5" }, "sft": { "method": "synthetic+openbook", "files": [ "sft/checked/mix.train.jsonl" ], "conversations": 1811163, "epochs": 1, "base_step": 3605, "init_lr_frac": 0.5, "mix": { "weights": { "synthetic": 0.55, "openbook": 0.25, "headings": 0.2 }, "anchor": "synthetic", "seed": 0, "sources": { "synthetic": { "available": 996393, "used": 996393 }, "openbook": { "available": 919976, "used": 452905 }, "headings": { "available": 521532, "used": 362324 } }, "shares": { "synthetic": 0.55, "openbook": 0.25, "headings": 0.2 }, "conversations": 1811622 }, "contamination": { "bench_files": { "arc_easy_test.json": { "sha256": "76e55faa15bce1b6ef26212c0c67a242a39952088d3b9e1a4d6071ea33c15190", "items": 2376 }, "cryptomc.json": { "sha256": "c81b3367d654f1e207d2dcbb48f490ff5feb496e2d08a049d39fb8aa8de35f87", "items": 150 }, "cryptoterms.json": { "sha256": "918b257c12deac8818cebf271187eedd88f026f875908b5a2e6987434b991f90", "items": 150 }, "ctslang.json": { "sha256": "d967e9196888027a4055592381892fca6185ee3a41b70be6a7351fa06d0bfc70", "items": 150 }, "projectid.json": { "sha256": "6feec798ea7269d2931fba111c48b33af2030a885e4c589235072ca6e488f51e", "items": 200 }, "projectid_xl.json": { "sha256": "859b5fc0d6a197bdf38e09e980d4577bd32f321e4fb3aec674725b6e4ed9a709", "items": 2593 }, "qa.json": { "sha256": "d5f699bdec9dca94e355332380adf317ef497784a56d3d9277ca6c0f993af9c1", "items": 100 } }, "skipped_files": [ "._arc_easy_test.json", "._cryptomc.json", "._cryptoterms.json", "._ctslang.json", "._fresh_pages.json", "._fresh_pages_log.json", "._grade-0.json", "._grade-1.json", "._grade-2.json", "._grade-3.json", "._grade-4.json", "._grade-key.json", "._grade2-a.json", "._grade2-b.json", "._grade2-c.json", "._grade2-grades.json", "._grade2-key.json", "._grade3-a.json", "._grade3-b.json", "._grade3-c.json", "._grade3-grades.json", "._grade3-key.json", "._models.json", "._projectid.json", "._projectid_bench_protocols.json", "._projectid_defillama_raw.json", "._projectid_meta.json", "._projectid_results.json", "._projectid_results_detail.json", "._projectid_split.json", "._projectid_xl.json", "._projectid_xl_meta.json", "._qa.json", "._qa_answers.json", "._report_q31.json", "._results.json", "._results_ctslang.json", "._results_ctslang_detail.json", "._results_detail.json", "._results_domain.json", "._results_domain_detail.json", "._smoke.json", "._web_snapshot.json", "fresh_pages.json", "fresh_pages_log.json", "grade-0.json", "grade-1.json", "grade-2.json", "grade-3.json", "grade-4.json", "grade-key.json", "grade2-a.json", "grade2-b.json", "grade2-c.json", "grade2-grades.json", "grade2-key.json", "grade3-a.json", "grade3-b.json", "grade3-c.json", "grade3-grades.json", "grade3-key.json", "models.json", "projectid_bench_protocols.json", "projectid_defillama_raw.json", "projectid_meta.json", "projectid_results.json", "projectid_results_detail.json", "projectid_split.json", "projectid_xl_meta.json", "qa_answers.json", "report_q31.json", "results.json", "results_ctslang.json", "results_ctslang_detail.json", "results_detail.json", "results_domain.json", "results_domain_detail.json", "smoke.json", "web_snapshot.json" ], "criteria": { "ngram13": 13, "ngram8": 8, "ngram8_share": 0.7, "question_min_words_in_text": 5, "answer_within_words": 100 }, "dropped": { "mix.train.jsonl": 459, "synthetic.val.jsonl": 5, "openbook.val.jsonl": 5, "headings.val.jsonl": 0 }, "checked": { "mix.train.jsonl": 1811622, "synthetic.val.jsonl": 20337, "openbook.val.jsonl": 18826, "headings.val.jsonl": 12012 }, "by_criterion": { "mix.train.jsonl": { "question": 451, "ngram8_70": 20, "question_answer": 19, "ngram13": 2 }, "synthetic.val.jsonl": { "question": 5 }, "openbook.val.jsonl": { "question": 5 } }, "by_bench": { "mix.train.jsonl": { "qa.json": 402, "cryptomc.json": 67, "projectid_xl.json": 10 }, "synthetic.val.jsonl": { "qa.json": 5 }, "openbook.val.jsonl": { "qa.json": 5 } }, "bench_fingerprint": "656b4d7aa50c7d85e43d3915aae7dce568d54be7842219a97db529bbc349c125" }, "synth": { "coverage": { "eligible": 188828, "answered": 146686, "with_pairs": 139073, "answered_share": 0.7768, "pairs_share": 0.7365 }, "gate": 0.75, "teacher": "a general-purpose LLM" } } }