{ "schema_version": 1, "status": "ADVANCE", "selected_label": "J69_s3598", "checkpoint_assessments": [ { "label": "J69_s0900", "step": 900, "all_2724_lexemes_scored": true, "lexical": { "accepted_exact_count": 1289, "accepted_exact_percent": 47.3201174743025, "empty_outputs": 0, "most_common_predictions": [ [ "wunjan", 31 ], [ "kurri-kurri", 26 ], [ "julbin", 22 ], [ "julbiji", 20 ], [ "juku", 20 ], [ "julbi", 17 ], [ "kurriyala", 17 ], [ "bakay", 17 ], [ "juka", 16 ], [ "walbul-walbul", 15 ], [ "jikal", 15 ], [ "mulkal-mulkal", 14 ], [ "kanunjul", 14 ], [ "bangkarr", 13 ], [ "kurrul", 13 ], [ "julaji", 13 ], [ "wuyumbu", 12 ], [ "nganyi", 11 ], [ "kalbarr", 11 ], [ "mara-waban-wabaji", 10 ] ], "passes_confidence_adjusted_gate": false, "prediction_file": "/workspace/v24.3-joint-lexeme-dose29-rerun1/predictions/J69_s0900.curated.json", "prediction_sha256": "8f7bb5dfac8ba9071d54e316ccc91306d0adb0c92f782772bc48eb45f2e3662e", "rows": 2724, "selected_exact_count": 1262, "source_copy_outputs": 67, "strata": { "identity_mapping": { "false": { "accepted_exact_count": 1232, "accepted_exact_percent": 46.194225721784775, "rows": 2667 }, "true": { "accepted_exact_count": 57, "accepted_exact_percent": 100.0, "rows": 57 } }, "multiple_accepted_references": { "false": { "accepted_exact_count": 1037, "accepted_exact_percent": 46.129893238434164, "rows": 2248 }, "true": { "accepted_exact_count": 252, "accepted_exact_percent": 52.94117647058823, "rows": 476 } }, "part_of_speech": { "adjective": { "accepted_exact_count": 167, "accepted_exact_percent": 43.717277486910994, "rows": 382 }, "adverb": { "accepted_exact_count": 35, "accepted_exact_percent": 47.945205479452056, "rows": 73 }, "demonstrative": { "accepted_exact_count": 8, "accepted_exact_percent": 80.0, "rows": 10 }, "interjection": { "accepted_exact_count": 4, "accepted_exact_percent": 30.76923076923077, "rows": 13 }, "interrogative": { "accepted_exact_count": 12, "accepted_exact_percent": 80.0, "rows": 15 }, "intransitive-verb": { "accepted_exact_count": 146, "accepted_exact_percent": 52.707581227436826, "rows": 277 }, "locational": { "accepted_exact_count": 21, "accepted_exact_percent": 72.41379310344827, "rows": 29 }, "noun": { "accepted_exact_count": 705, "accepted_exact_percent": 43.62623762376238, "rows": 1616 }, "particle": { "accepted_exact_count": 11, "accepted_exact_percent": 68.75, "rows": 16 }, "personal-pronoun": { "accepted_exact_count": 11, "accepted_exact_percent": 64.70588235294117, "rows": 17 }, "temporal": { "accepted_exact_count": 28, "accepted_exact_percent": 80.0, "rows": 35 }, "transitive-verb": { "accepted_exact_count": 221, "accepted_exact_percent": 58.776595744680854, "rows": 376 } }, "source_word_count": { "1": { "accepted_exact_count": 884, "accepted_exact_percent": 56.557901471529114, "rows": 1563 }, "2": { "accepted_exact_count": 290, "accepted_exact_percent": 36.523929471032744, "rows": 794 }, "3-5": { "accepted_exact_count": 106, "accepted_exact_percent": 31.085043988269796, "rows": 341 }, "6+": { "accepted_exact_count": 9, "accepted_exact_percent": 34.61538461538461, "rows": 26 } } }, "wilson_95": { "high": 0.4919752520725862, "low": 0.4545025758349349 } }, "retention": { "checks": { "all_outputs_nonempty": true, "elder_diagnostic_chrf_noninferior": true, "elder_diagnostic_empty_zero": true, "elder_diagnostic_loops_not_increased": true, "lexical_exact_gain_over_control": true, "natural_dev_text36_chrf_noninferior": true, "natural_dev_text36_empty_zero": true, "natural_dev_text36_loops_not_increased": true, "synthetic_dev_chrf_noninferior": true, "synthetic_dev_empty_zero": true, "synthetic_dev_loops_not_increased": true, "usage_diagnostic_chrf_noninferior": true, "usage_diagnostic_empty_zero": true, "usage_diagnostic_loops_not_increased": true }, "dose": 1000000, "eligible": true, "lexical_exact_gain_over_control": 20, "lexical_exact_paired_difference": { "descriptive_row_bootstrap_95_high": 11.447811447811448, "descriptive_row_bootstrap_95_low": 2.356902356902357, "percentage_points": 6.7340067340067336 }, "mean_retention_chrf_delta": 2.231445078930668, "minimum_retention_chrf_delta": 0.439048848140132, "retention_chrf_delta_from_stronger_control": { "elder_diagnostic": 0.439048848140132, "natural_dev_text36": 3.1980733149968437, "synthetic_dev": 4.512079964857236, "usage_diagnostic": 0.776578187728461 } }, "failure_categories": { "accepted_with_extra_material": 17, "exact_alternative_reference": 27, "exact_selected": 1262, "known_target_for_other_prompt": 1134, "near_orthographic": 32, "notation_or_hyphen_mismatch": 1, "other": 241, "source_copy": 10 }, "lexical_gate_pass": false, "sentence_retention_gate_pass": true, "joint_gate_pass": false }, { "label": "J69_s1800", "step": 1800, "all_2724_lexemes_scored": true, "lexical": { "accepted_exact_count": 2321, "accepted_exact_percent": 85.20558002936858, "empty_outputs": 0, "most_common_predictions": [ [ "kalbarr", 8 ], [ "jaka", 7 ], [ "ngurma", 7 ], [ "dirka", 7 ], [ "badi", 7 ], [ "wawu", 7 ], [ "balu", 6 ], [ "dumbul", 6 ], [ "wawurr-wawurr-bungal", 6 ], [ "dijarr", 6 ], [ "jarramali", 6 ], [ "ngamba", 5 ], [ "kanunjul", 5 ], [ "kilki", 5 ], [ "mara-waban-wabaji", 5 ], [ "muyul-muyul", 5 ], [ "mulkal-mulkal", 5 ], [ "dingal", 5 ], [ "ngamu", 5 ], [ "wunjan", 5 ] ], "passes_confidence_adjusted_gate": true, "prediction_file": "/workspace/v24.3-joint-lexeme-dose29-rerun1/predictions/J69_s1800.curated.json", "prediction_sha256": "ca4d697430c2c584ef160f38d35f48a863b8d94e077e55b4ce1726a837e73da0", "rows": 2724, "selected_exact_count": 2307, "source_copy_outputs": 61, "strata": { "identity_mapping": { "false": { "accepted_exact_count": 2264, "accepted_exact_percent": 84.8893888263967, "rows": 2667 }, "true": { "accepted_exact_count": 57, "accepted_exact_percent": 100.0, "rows": 57 } }, "multiple_accepted_references": { "false": { "accepted_exact_count": 1906, "accepted_exact_percent": 84.7864768683274, "rows": 2248 }, "true": { "accepted_exact_count": 415, "accepted_exact_percent": 87.18487394957984, "rows": 476 } }, "part_of_speech": { "adjective": { "accepted_exact_count": 335, "accepted_exact_percent": 87.69633507853403, "rows": 382 }, "adverb": { "accepted_exact_count": 70, "accepted_exact_percent": 95.89041095890411, "rows": 73 }, "demonstrative": { "accepted_exact_count": 7, "accepted_exact_percent": 70.0, "rows": 10 }, "interjection": { "accepted_exact_count": 11, "accepted_exact_percent": 84.61538461538461, "rows": 13 }, "interrogative": { "accepted_exact_count": 15, "accepted_exact_percent": 100.0, "rows": 15 }, "intransitive-verb": { "accepted_exact_count": 235, "accepted_exact_percent": 84.83754512635379, "rows": 277 }, "locational": { "accepted_exact_count": 27, "accepted_exact_percent": 93.10344827586206, "rows": 29 }, "noun": { "accepted_exact_count": 1350, "accepted_exact_percent": 83.53960396039604, "rows": 1616 }, "particle": { "accepted_exact_count": 16, "accepted_exact_percent": 100.0, "rows": 16 }, "personal-pronoun": { "accepted_exact_count": 14, "accepted_exact_percent": 82.3529411764706, "rows": 17 }, "temporal": { "accepted_exact_count": 30, "accepted_exact_percent": 85.71428571428571, "rows": 35 }, "transitive-verb": { "accepted_exact_count": 332, "accepted_exact_percent": 88.29787234042553, "rows": 376 } }, "source_word_count": { "1": { "accepted_exact_count": 1417, "accepted_exact_percent": 90.65898912348048, "rows": 1563 }, "2": { "accepted_exact_count": 618, "accepted_exact_percent": 77.8337531486146, "rows": 794 }, "3-5": { "accepted_exact_count": 271, "accepted_exact_percent": 79.47214076246334, "rows": 341 }, "6+": { "accepted_exact_count": 15, "accepted_exact_percent": 57.69230769230769, "rows": 26 } } }, "wilson_95": { "high": 0.8648928396679225, "low": 0.8382272019305337 } }, "retention": { "checks": { "all_outputs_nonempty": true, "elder_diagnostic_chrf_noninferior": true, "elder_diagnostic_empty_zero": true, "elder_diagnostic_loops_not_increased": true, "lexical_exact_gain_over_control": true, "natural_dev_text36_chrf_noninferior": true, "natural_dev_text36_empty_zero": true, "natural_dev_text36_loops_not_increased": true, "synthetic_dev_chrf_noninferior": true, "synthetic_dev_empty_zero": true, "synthetic_dev_loops_not_increased": true, "usage_diagnostic_chrf_noninferior": true, "usage_diagnostic_empty_zero": true, "usage_diagnostic_loops_not_increased": true }, "dose": 1000000, "eligible": true, "lexical_exact_gain_over_control": 40, "lexical_exact_paired_difference": { "descriptive_row_bootstrap_95_high": 18.855218855218855, "descriptive_row_bootstrap_95_low": 8.417508417508419, "percentage_points": 13.468013468013467 }, "mean_retention_chrf_delta": 4.905562167460687, "minimum_retention_chrf_delta": 0.945519864198527, "retention_chrf_delta_from_stronger_control": { "elder_diagnostic": 0.945519864198527, "natural_dev_text36": 5.100579522145747, "synthetic_dev": 8.62012268644063, "usage_diagnostic": 4.956026597057843 } }, "failure_categories": { "accepted_with_extra_material": 3, "exact_alternative_reference": 14, "exact_selected": 2307, "known_target_for_other_prompt": 324, "near_orthographic": 13, "other": 59, "source_copy": 4 }, "lexical_gate_pass": true, "sentence_retention_gate_pass": true, "joint_gate_pass": true }, { "label": "J69_s2700", "step": 2700, "all_2724_lexemes_scored": true, "lexical": { "accepted_exact_count": 2633, "accepted_exact_percent": 96.65932452276064, "empty_outputs": 0, "most_common_predictions": [ [ "ngurma", 7 ], [ "wawurr-wawurr-bungal", 6 ], [ "dirrmbay", 6 ], [ "dirka", 6 ], [ "wawu", 6 ], [ "mambarril", 6 ], [ "balu", 5 ], [ "mara-waban-wabaji", 5 ], [ "mulkal-mulkal", 5 ], [ "wunjan", 5 ], [ "marra-warril", 5 ], [ "dandi", 5 ], [ "jaba", 5 ], [ "jarramali", 5 ], [ "nyarril", 5 ], [ "jirra-kangal", 5 ], [ "ngamba", 4 ], [ "kaka", 4 ], [ "jaba-jaba", 4 ], [ "kalngar", 4 ] ], "passes_confidence_adjusted_gate": true, "prediction_file": "/workspace/v24.3-joint-lexeme-dose29-rerun1/predictions/J69_s2700.curated.json", "prediction_sha256": "764ec11ba329a3146a66f2e546e66ff0d14e948d8ce62a310a12cc9f4ce849b0", "rows": 2724, "selected_exact_count": 2629, "source_copy_outputs": 57, "strata": { "identity_mapping": { "false": { "accepted_exact_count": 2576, "accepted_exact_percent": 96.58792650918635, "rows": 2667 }, "true": { "accepted_exact_count": 57, "accepted_exact_percent": 100.0, "rows": 57 } }, "multiple_accepted_references": { "false": { "accepted_exact_count": 2173, "accepted_exact_percent": 96.66370106761566, "rows": 2248 }, "true": { "accepted_exact_count": 460, "accepted_exact_percent": 96.63865546218487, "rows": 476 } }, "part_of_speech": { "adjective": { "accepted_exact_count": 374, "accepted_exact_percent": 97.90575916230367, "rows": 382 }, "adverb": { "accepted_exact_count": 73, "accepted_exact_percent": 100.0, "rows": 73 }, "demonstrative": { "accepted_exact_count": 10, "accepted_exact_percent": 100.0, "rows": 10 }, "interjection": { "accepted_exact_count": 13, "accepted_exact_percent": 100.0, "rows": 13 }, "interrogative": { "accepted_exact_count": 15, "accepted_exact_percent": 100.0, "rows": 15 }, "intransitive-verb": { "accepted_exact_count": 268, "accepted_exact_percent": 96.75090252707581, "rows": 277 }, "locational": { "accepted_exact_count": 29, "accepted_exact_percent": 100.0, "rows": 29 }, "noun": { "accepted_exact_count": 1551, "accepted_exact_percent": 95.97772277227723, "rows": 1616 }, "particle": { "accepted_exact_count": 16, "accepted_exact_percent": 100.0, "rows": 16 }, "personal-pronoun": { "accepted_exact_count": 17, "accepted_exact_percent": 100.0, "rows": 17 }, "temporal": { "accepted_exact_count": 34, "accepted_exact_percent": 97.14285714285714, "rows": 35 }, "transitive-verb": { "accepted_exact_count": 366, "accepted_exact_percent": 97.34042553191489, "rows": 376 } }, "source_word_count": { "1": { "accepted_exact_count": 1529, "accepted_exact_percent": 97.82469609724888, "rows": 1563 }, "2": { "accepted_exact_count": 761, "accepted_exact_percent": 95.84382871536523, "rows": 794 }, "3-5": { "accepted_exact_count": 320, "accepted_exact_percent": 93.841642228739, "rows": 341 }, "6+": { "accepted_exact_count": 23, "accepted_exact_percent": 88.46153846153847, "rows": 26 } } }, "wilson_95": { "high": 0.9727114852502817, "low": 0.9591608535105504 } }, "retention": { "checks": { "all_outputs_nonempty": true, "elder_diagnostic_chrf_noninferior": true, "elder_diagnostic_empty_zero": true, "elder_diagnostic_loops_not_increased": true, "lexical_exact_gain_over_control": true, "natural_dev_text36_chrf_noninferior": true, "natural_dev_text36_empty_zero": true, "natural_dev_text36_loops_not_increased": true, "synthetic_dev_chrf_noninferior": true, "synthetic_dev_empty_zero": true, "synthetic_dev_loops_not_increased": true, "usage_diagnostic_chrf_noninferior": true, "usage_diagnostic_empty_zero": false, "usage_diagnostic_loops_not_increased": true }, "dose": 1000000, "eligible": false, "lexical_exact_gain_over_control": 45, "lexical_exact_paired_difference": { "descriptive_row_bootstrap_95_high": 20.2020202020202, "descriptive_row_bootstrap_95_low": 10.1010101010101, "percentage_points": 15.151515151515152 }, "mean_retention_chrf_delta": 5.776385216810932, "minimum_retention_chrf_delta": 1.0722849961428835, "retention_chrf_delta_from_stronger_control": { "elder_diagnostic": 1.0722849961428835, "natural_dev_text36": 5.999971872055813, "synthetic_dev": 9.702506484001347, "usage_diagnostic": 6.330777515043685 } }, "failure_categories": { "exact_alternative_reference": 4, "exact_selected": 2629, "known_target_for_other_prompt": 84, "near_orthographic": 4, "other": 3 }, "lexical_gate_pass": true, "sentence_retention_gate_pass": false, "joint_gate_pass": false }, { "label": "J69_s3598", "step": 3598, "all_2724_lexemes_scored": true, "lexical": { "accepted_exact_count": 2696, "accepted_exact_percent": 98.97209985315712, "empty_outputs": 0, "most_common_predictions": [ [ "ngurma", 6 ], [ "wawurr-wawurr-bungal", 6 ], [ "dirrmbay", 6 ], [ "dirka", 6 ], [ "wawu", 6 ], [ "balu", 5 ], [ "mara-waban-wabaji", 5 ], [ "mulkal-mulkal", 5 ], [ "wunjan", 5 ], [ "marra-warril", 5 ], [ "mambarril", 5 ], [ "dandi", 5 ], [ "jaba", 5 ], [ "jirra-kangal", 5 ], [ "ngamba", 4 ], [ "kaka", 4 ], [ "jaba-jaba", 4 ], [ "kalngar", 4 ], [ "kilki", 4 ], [ "yungal", 4 ] ], "passes_confidence_adjusted_gate": true, "prediction_file": "/workspace/v24.3-joint-lexeme-dose29-rerun1/predictions/J69_s3598.curated.json", "prediction_sha256": "a9ea6b9cc1e3576d0cb9ffe6eb7f8325a0778e0c652f739cf62848d76239c5b4", "rows": 2724, "selected_exact_count": 2694, "source_copy_outputs": 57, "strata": { "identity_mapping": { "false": { "accepted_exact_count": 2639, "accepted_exact_percent": 98.9501312335958, "rows": 2667 }, "true": { "accepted_exact_count": 57, "accepted_exact_percent": 100.0, "rows": 57 } }, "multiple_accepted_references": { "false": { "accepted_exact_count": 2225, "accepted_exact_percent": 98.97686832740213, "rows": 2248 }, "true": { "accepted_exact_count": 471, "accepted_exact_percent": 98.94957983193277, "rows": 476 } }, "part_of_speech": { "adjective": { "accepted_exact_count": 381, "accepted_exact_percent": 99.73821989528795, "rows": 382 }, "adverb": { "accepted_exact_count": 73, "accepted_exact_percent": 100.0, "rows": 73 }, "demonstrative": { "accepted_exact_count": 10, "accepted_exact_percent": 100.0, "rows": 10 }, "interjection": { "accepted_exact_count": 13, "accepted_exact_percent": 100.0, "rows": 13 }, "interrogative": { "accepted_exact_count": 15, "accepted_exact_percent": 100.0, "rows": 15 }, "intransitive-verb": { "accepted_exact_count": 274, "accepted_exact_percent": 98.91696750902527, "rows": 277 }, "locational": { "accepted_exact_count": 29, "accepted_exact_percent": 100.0, "rows": 29 }, "noun": { "accepted_exact_count": 1595, "accepted_exact_percent": 98.70049504950495, "rows": 1616 }, "particle": { "accepted_exact_count": 16, "accepted_exact_percent": 100.0, "rows": 16 }, "personal-pronoun": { "accepted_exact_count": 17, "accepted_exact_percent": 100.0, "rows": 17 }, "temporal": { "accepted_exact_count": 33, "accepted_exact_percent": 94.28571428571429, "rows": 35 }, "transitive-verb": { "accepted_exact_count": 374, "accepted_exact_percent": 99.46808510638297, "rows": 376 } }, "source_word_count": { "1": { "accepted_exact_count": 1555, "accepted_exact_percent": 99.48816378758796, "rows": 1563 }, "2": { "accepted_exact_count": 784, "accepted_exact_percent": 98.74055415617129, "rows": 794 }, "3-5": { "accepted_exact_count": 333, "accepted_exact_percent": 97.65395894428153, "rows": 341 }, "6+": { "accepted_exact_count": 24, "accepted_exact_percent": 92.3076923076923, "rows": 26 } } }, "wilson_95": { "high": 0.9928787071997702, "low": 0.9851839992573108 } }, "retention": { "checks": { "all_outputs_nonempty": true, "elder_diagnostic_chrf_noninferior": true, "elder_diagnostic_empty_zero": true, "elder_diagnostic_loops_not_increased": true, "lexical_exact_gain_over_control": true, "natural_dev_text36_chrf_noninferior": true, "natural_dev_text36_empty_zero": true, "natural_dev_text36_loops_not_increased": true, "synthetic_dev_chrf_noninferior": true, "synthetic_dev_empty_zero": true, "synthetic_dev_loops_not_increased": true, "usage_diagnostic_chrf_noninferior": true, "usage_diagnostic_empty_zero": true, "usage_diagnostic_loops_not_increased": true }, "dose": 1000000, "eligible": true, "lexical_exact_gain_over_control": 43, "lexical_exact_paired_difference": { "descriptive_row_bootstrap_95_high": 19.865319865319865, "descriptive_row_bootstrap_95_low": 9.090909090909092, "percentage_points": 14.47811447811448 }, "mean_retention_chrf_delta": 5.370022937107108, "minimum_retention_chrf_delta": 0.6656234052750634, "retention_chrf_delta_from_stronger_control": { "elder_diagnostic": 0.6656234052750634, "natural_dev_text36": 5.594655686933461, "synthetic_dev": 10.462328446429353, "usage_diagnostic": 4.7574842097905545 } }, "failure_categories": { "exact_alternative_reference": 2, "exact_selected": 2694, "known_target_for_other_prompt": 25, "near_orthographic": 2, "other": 1 }, "lexical_gate_pass": true, "sentence_retention_gate_pass": true, "joint_gate_pass": true } ], "trajectory_summary": { "adjacent_transitions": { "B0->J69_s0900": { "gain": 1082, "loss": 19, "stable_exact": 207, "stable_failure": 1416 }, "J69_s0900->J69_s1800": { "gain": 1084, "loss": 52, "stable_exact": 1237, "stable_failure": 351 }, "J69_s1800->J69_s2700": { "gain": 334, "loss": 22, "stable_exact": 2299, "stable_failure": 69 }, "J69_s2700->J69_s3598": { "gain": 66, "loss": 3, "stable_exact": 2630, "stable_failure": 25 } }, "analysis_kind": "aligned_closed_set_lexeme_checkpoint_trajectories", "baseline_label": "B0", "candidate_order": [ "B0", "J69_s0900", "J69_s1800", "J69_s2700", "J69_s3598" ], "checkpoint_error_categories": { "B0": { "accepted_with_extra_material": 18, "empty": 1, "exact_alternative_reference": 29, "exact_selected": 197, "known_target_for_other_prompt": 1389, "near_orthographic": 43, "notation_or_hyphen_mismatch": 4, "other": 984, "source_copy": 59 }, "J69_s0900": { "accepted_with_extra_material": 17, "exact_alternative_reference": 27, "exact_selected": 1262, "known_target_for_other_prompt": 1134, "near_orthographic": 32, "notation_or_hyphen_mismatch": 1, "other": 241, "source_copy": 10 }, "J69_s1800": { "accepted_with_extra_material": 3, "exact_alternative_reference": 14, "exact_selected": 2307, "known_target_for_other_prompt": 324, "near_orthographic": 13, "other": 59, "source_copy": 4 }, "J69_s2700": { "exact_alternative_reference": 4, "exact_selected": 2629, "known_target_for_other_prompt": 84, "near_orthographic": 4, "other": 3 }, "J69_s3598": { "exact_alternative_reference": 2, "exact_selected": 2694, "known_target_for_other_prompt": 25, "near_orthographic": 2, "other": 1 } }, "created_at": "2026-07-15T17:46:59.809890+00:00", "interpretation": "All rows are training-overlapping closed-set reconstruction diagnostics. Trajectories distinguish optimization persistence and regression; they do not establish unseen lexical or sentence competence.", "row_analysis_file": "/workspace/v24.3-joint-lexeme-dose29-rerun1/analysis/all-checkpoint-row-analysis.jsonl", "row_analysis_sha256": "9e45173dfc4608bc900c7d0dbee2f2d31c441ddde839a50fc417d0065bf95059", "rows": 2724, "schema_version": 1, "trajectory_classes": { "exact_at_every_checkpoint": 1230, "learned_and_retained": 1396, "never_exact": 21, "oscillating": 70, "regressed": 7 }, "treatment_labels": [ "J69_s0900", "J69_s1800", "J69_s2700", "J69_s3598" ] }, "observed_global_step": 3598, "actual_training_exposure": { "examples": 115136, "source_tokens": 1268518, "target_tokens": 1122677, "non_padding_tokens": 2391195, "by_task": { "dictionary_lexeme": { "examples": 78996, "source_tokens": 441902, "target_tokens": 384569, "non_padding_tokens": 826471 }, "synthetic_academic_parallel": { "examples": 33284, "source_tokens": 713552, "target_tokens": 704692, "non_padding_tokens": 1418244 }, "usage_example": { "examples": 2856, "source_tokens": 113064, "target_tokens": 33416, "non_padding_tokens": 146480 } }, "unique_rows_seen": 115136, "presentations_per_seen_row": { "minimum": 1, "maximum": 1, "mean": 1.0 }, "accounting_scope": "current trainer process", "world_size": 1 }, "interpretation": "Closed-set lexical reconstruction uses training-overlapping dictionary records. Passing it does not establish unseen lexical generalization or sentence translation reliability.", "promotion_or_deployment_authorized": false }