{ "schema_version": "geml-detailed-results-ledger-v1", "as_of": "2026-08-02", "scope": "Authenticated Goals 1-5 outputs plus authenticated Phase-B GPU outputs. Raw rows, checkpoints, logs, and archives are excluded.", "provenance": { "repository_snapshot": "ANON-CURRENT-SNAPSHOT", "repository_snapshot_label": "ANON-CURRENT-SNAPSHOT", "artifact_root": "/GEML_artifacts", "phase_b_package": "Phase_B_GPU_results/phase-b-20260731-anonymous", "phase_b_source_snapshot": "ANON-PHASE-B-SNAPSHOT" }, "goal1": { "status": "complete", "source": "1-8_source_expression_corpus_250k/manifests/corpus.manifest.json; 1-8_source_expression_corpus_250k/qa.report.json; 1-8_source_expression_corpus_250k/stage.result.json; 1-8_source_expression_corpus_250k/run.metadata.json", "schema": "geml-corpus-v1", "corpus_id": "geml-goal1-final", "generator_seed": 20260721, "generator_snapshot": "ANON-G1-SNAPSHOT", "config_hash": "2a8381a53fd4a69473d2ddf0fda3860885c80976f5fdd9fca465bfede8223538", "package_versions": {"geml": "0.1.0", "mpmath": "1.3.0", "numpy": "2.5.1", "psutil": "7.2.2", "pyarrow": "25.0.0", "pydantic": "2.13.4", "sympy": "1.14.0"}, "platform": "Windows-11-10.0.26200-SP0", "python": "3.12.13", "accounting": {"attempted": 286413, "generated": 286413, "accepted": 250000, "finalized": 250000, "duplicates": 35768, "identity_conflicts": 0, "triviality_rejections": 645, "internal_triviality_retries": 500, "unsupported": 0, "policy_rejections": 0, "parse_failures": 0, "ast_validation_failures": 0, "display_failures": 0, "latex_failures": 0, "roundtrip_failures": 0, "storage_failures": 0, "acceptance_rate": 0.8728654076456027}, "splits": {"train": 175000, "validation": 25000, "test_iid": 25000, "test_ood": 25000}, "domains": {"nonzero_real": 68345, "positive_real": 84843, "safe_real": 96812}, "families": {"algebraic_core": 70000, "exp_log": 40000, "mixed_elementary": 35000, "ood_stress": 25000, "powers_division_rationals": 40000, "trig_hyperbolic": 40000}, "variables": {"1": 68934, "2": 61244, "3": 46743, "4": 34511, "5": 24095, "6": 14473}, "depth_distributions": { "target_source_depth": {"1": 373, "2": 5277, "3": 24644, "4": 47385, "5": 52127, "6": 44935, "7": 28770, "8": 20555, "9": 11913, "10": 7038, "11": 3836, "12": 3147}, "actual_ast_depth": {"1": 213, "2": 2870, "3": 11603, "4": 25525, "5": 38367, "6": 42363, "7": 39623, "8": 32685, "9": 24053, "10": 15090, "11": 8942, "12": 5310, "13": 2479, "14": 728, "15": 124, "16": 24, "17": 1}, "actual_minus_target_depth": {"0": 72958, "1": 102154, "2": 55896, "3": 16332, "4": 2435, "5": 208, "6": 17}, "summary": {"target_source_depth_min": 1, "target_source_depth_max": 12, "target_source_depth_mean": 5.712124, "actual_ast_depth_min": 1, "actual_ast_depth_max": 17, "actual_ast_depth_mean": 6.80742, "actual_minus_target_depth_mean": 1.095296} }, "size_distributions_summary": {"target_source_ast_size_min": 2, "target_source_ast_size_max": 128, "target_source_ast_size_mean": 31.114768, "target_source_ast_size_median": 22, "actual_ast_node_count_min": 2, "actual_ast_node_count_max": 213, "actual_ast_node_count_mean": 40.791744, "actual_ast_node_count_median": 29, "actual_minus_target_size_min": 0, "actual_minus_target_size_max": 86, "actual_minus_target_size_mean": 9.676976, "actual_minus_target_size_median": 7}, "constant_leaf_counts": {"integer": 2257365, "one": 255516, "rational": 447949}, "operator_usage": {"add": 986611, "cos": 68080, "cosh": 66083, "divide": 253949, "exp": 448192, "integer": 730763, "log": 196656, "multiply": 996331, "negate": 735056, "one": 204968, "power": 82926, "rational": 447949, "sin": 67485, "sinh": 48996, "subtract": 588145, "symbol": 1774282, "tan": 28648, "tanh": 53572}, "policy": {"all_trig_operators_covered": true, "approved_operator_domain_check": true, "blanket_log_exp_detected": false, "certified_log_arguments": 196656, "certified_tan_arguments": 28648, "lowered_reciprocal_candidates": 257698, "negative_power_arguments": 292783, "log_argument_policy": "positive_expression_grammar", "tan_argument_policy": "closed_unit_interval_structural_grammar"}, "triviality": {"record_limits": {"constant_only_subtrees": 175000, "exp_log": 87500, "log_exp": 100000, "log_one": 20000, "multiplication_by_one": 50000}, "selected_record_counts": {"constant_only_subtrees": 159469, "exp_log": 34685, "log_exp": 42093, "log_one": 6400, "multiplication_by_one": 50000}, "record_rates": {"constant_only_subtrees": 0.637876, "exp_log": 0.13874, "log_exp": 0.168372, "log_one": 0.0256, "multiplication_by_one": 0.2}, "rate_cap_violations": {}}, "qa": {"loaded_rows": 250000, "manifest_rows": 250000, "validated_rows": 250000, "validated_shards": 10, "unique_authoritative_srepr": 250000, "unique_expression_ids": 250000, "unique_structural_identities": 250000, "cross_split_expression_ids": 0, "duplicate_expression_id_occurrences": 0, "duplicate_srepr_occurrences": 0, "duplicate_structural_identity_occurrences": 0, "latex_parser_unavailable_samples": 64, "latex_roundtrip_failures": 0}, "timing": {"manifest_elapsed_seconds": 620.2928233999992, "qa_inclusive_elapsed_seconds": 899.2492879000201, "accepted_throughput_rows_per_second": 403.0354544966034, "qa_inclusive_accepted_throughput_rows_per_second": 278.00967247226265, "generation_throughput_rows_per_second": 461.7383745149426, "qa_inclusive_generation_throughput_rows_per_second": 318.50233728719263, "manifest_peak_rss_bytes": 2193186816, "qa_inclusive_peak_rss_bytes": 3044659200, "timings_seconds": {"ast": 113.64893130990095, "display": 25.29359439588734, "latex": 23.915461296361173, "parse": 102.23721850299626, "qa": 264.13349199999357, "shard_write": 4.825599499978125, "split_assignment": 2.721126499993261}} }, "goal2": { "status": "complete", "source": "2-7_2-8_official_pure_eml_corpus/manifest.json; summary.json; docs/goals/goal2/GOAL2_EXPANSION_STUDY.md", "schema": "geml-goal2-summary-v1", "compiler_mode": "official_v4", "input_label": "final_goal1_corpus_250k", "accounting": {"processed": 250000, "success": 250000, "conversion_failures": 0, "semantic_selected": 280, "semantic_materialized_and_audited": 273, "semantic_passed": 203, "semantic_mismatch": 3, "semantic_nonfinite": 45, "semantic_overflow": 22, "node_limit_before_materialization": 7, "semantic_not_selected": 249720}, "raw_pure_eml_alpha": {"median": 40.6602, "mean": 952.1371, "p99_approx": 10448.6, "rows_below_preregistered_threshold_range_1_29_to_1_50": 0}, "timing": {"elapsed_seconds": 2557.2617339999997, "processing_rows_per_second": 99.16671338859994, "aggregation_rows_per_second": 97.76081840827327, "peak_rss_bytes": 1617088512}, "interpretation": "The 273 materialized semantic-audit statuses are a subset audit, not conversion failures or a claim that unmaterialized rows are invalid." }, "goal3": { "status": "complete", "source": "3-1_to_3-8_graph_corpus_and_costs/summary.json; run.metadata.json; 3-7_3-8_goal3_analysis/metrics.summary.json", "compiler_mode": "official_v4", "construction": "direct_hashcons", "accounting": {"processed": 250000, "success": 250000, "failures": 0, "audit_gate_ready": true}, "timing": {"runtime_seconds": 1568.0315, "throughput_rows_per_second": 159.4356, "cumulative_throughput_rows_per_second": 162.7917, "peak_rss_bytes": 1444044800}, "ast_dag": {"max_indegree": 63, "reused_nodes_total": 857703, "reused_references_total": 3439858, "excess_references_total": 2582155, "mean_reused_nodes": 3.430812, "mean_reused_references": 13.759432, "mean_excess_references": 10.32862, "mean_reuse_depth": 4.6421511875, "sharing_concentration": 0.5698556037}, "eml_dag": {"max_indegree": 1939, "reused_nodes_total": 4532684, "reused_references_total": 69048387, "excess_references_total": 64515703, "mean_reused_nodes": 18.130736, "mean_reused_references": 276.193548, "mean_excess_references": 258.062812, "mean_reuse_depth": 27.7799859862, "sharing_concentration": 0.7978003017}, "ratios": {"raw_tree_alpha_mean": 952.1371252900141, "eml_dag_vs_ast_tree_mean": 8.334401271758448, "eml_dag_vs_ast_dag_mean": 10.474953890182578, "ast_compression_mean": 1.361707663394114, "eml_compression_mean": 39.37500771693145, "structurally_competitive_rows": 0, "best_remaining_ratio": "8/7"} }, "goal4": { "status": "complete", "source": "4-1_to_4-9_goal4_egraph_study/outputs/final/goal4/analysis/summary.json; failures.json; plot_data.json", "run_id": "9c26ec3036c45bd3bf24256d9a57fa4e1e48d016cd2cee446a47918667cf2536", "total_rows": 60000, "selected_expressions": 30000, "modes": { "safe_real": {"processed": 30000, "costed": 18210, "improved": 4349, "failed": 11790, "timeouts": 0, "improved_over_costed": 0.2388248215, "improved_over_processed": 0.1449666667, "mean_signed_node_improvement": 5.2496430533, "mean_relative_improvement": 0.0236117313, "failure_categories": {"unsupported_operator": 11566, "validation_failed": 224}, "failure_rate_permille": 393}, "positive_real_formal": {"processed": 30000, "costed": 18210, "improved": 5026, "failed": 11790, "timeouts": 0, "improved_over_costed": 0.2760021966, "improved_over_processed": 0.1675333333, "mean_signed_node_improvement": 5.7658429434, "mean_relative_improvement": 0.02848704193, "failure_categories": {"unsupported_operator": 11566, "validation_failed": 224}, "failure_rate_permille": 393} }, "failure_reasons": ["operator tanh/cosh/sinh/sin/tan/cos outside e-graph vocabulary", "no candidate passed independent validation (inconclusive candidates retained)"], "family_failure_counts_per_mode": {"algebraic_core": {"processed": 5884, "failed": 0}, "exp_log": {"processed": 5662, "failed": 134}, "mixed_elementary": {"processed": 5663, "failed": 5663}, "ood_stress": {"processed": 1107, "failed": 90}, "powers_division_rationals": {"processed": 5781, "failed": 0}, "trig_hyperbolic": {"processed": 5903, "failed": 5903}}, "validation_failures_per_mode": {"safe_real": 224, "positive_real_formal": 225}, "note": "The published failure categories are the same aggregate unsupported/validation categories for both modes; the one-row difference in validation counts is retained separately by the source report." }, "goal5": { "status": "complete", "source": "5-5_frequent_motif_mdl_sweep; 5-6_learned_motif_selection; 5-7_neural_egraph_candidate_ranker; 5-9_goals1_to_5_final_report/goal5.plot-data.json and goal5.summary.json", "frequent_motif": {"vocabulary_size": 1024, "motif_sizes": [2, 3, 4], "min_support": 32, "training_rows": 175000, "candidate_counts_by_size": {"1": 143, "2": 1984, "3": 35137, "4": 210653, "5": 734533, "6": 1847908, "7": 3688321, "8": 7187476}, "frequent_counts_by_size": {"1": 143, "2": 1259, "3": 6225, "4": 14137, "5": 28567, "6": 40586, "7": 79671, "8": 130853}, "train_mdl_bits": 2180887611, "validation_mdl_bits": 318470750, "test_iid": {"rows": 25000, "baseline_total_bits": 572524716, "total_mdl_bits": 317678264, "savings_bits": 254846452, "savings_fraction": 0.44512742398356125, "selected_occurrences": 95970, "reconstruction_failures": 0}, "test_ood": {"rows": 25000, "baseline_total_bits": 1453062331, "total_mdl_bits": 818099441, "savings_bits": 634962890, "savings_fraction": 0.43698255501745575, "selected_occurrences": 232143, "reconstruction_failures": 0}}, "learned_motif": {"vocabulary_size": 1024, "motif_sizes": [2, 3, 4], "min_support": 32, "candidate_source": 21621, "prefiltered_candidates": 4096, "training_examples": 4096, "training_seed": 20260723, "ridge_lambda": 100, "validation_mdl_bits": 325199669, "test_iid": {"total_mdl_bits": 324485346, "equal_budget_frequent_bits": 317678264, "random_median_bits": 368943015, "uncompressed_macro_bits": 572524716, "reconstruction_failures": 0}, "test_ood": {"total_mdl_bits": 821843999, "equal_budget_frequent_bits": 818099441, "uncompressed_macro_bits": 1453062331, "reconstruction_failures": 0}, "claim_vs_equal_budget_frequent": "null", "claim_vs_random_and_macro": "positive"}, "graph_tracks_250k": {"ast_dag": {"mean_nodes": 28.7112, "mean_edges": 38.03982, "total_nodes": 7177800, "total_edges": 9509955, "success": 250000, "failures": 0}, "pure_eml_dag": {"mean_nodes": 264.478876, "mean_edges": 521.541688, "total_nodes": 66119719, "total_edges": 130385422, "success": 250000, "failures": 0}, "macro_dag": {"mean_nodes": 28.7112, "mean_edges": 38.03982, "mean_mdl_bits": 26453.12424, "total_mdl_bits": 6613281060, "success": 250000, "failures": 0}, "frequent_motif_dag": {"mean_nodes": 19.268184, "mean_edges": 26.785816, "mean_mdl_bits": 14540.544264, "total_mdl_bits": 3635136066, "success": 250000, "failures": 0}, "learned_motif_dag": {"mean_nodes": 19.372924, "mean_edges": 26.9401, "mean_mdl_bits": 19620.3868533, "mean_mdl_denominator": 75000, "total_mdl_bits": 1471529014, "success": 250000, "failures": 0}}, "graph_tracks_30k": {"safe_egraph_eml_dag": {"rows": 30000, "costed": 18210, "mean_nodes_costed": 233.708566722, "failures": 11790}, "domain_egraph_eml_dag": {"rows": 30000, "costed": 18210, "mean_nodes_costed": 233.192366831, "failures": 11790}}, "neural_ranker_config": {"method": "geml-egraph-random-feature-ranker-v1", "hidden_units": 32, "target_transform": "log1p", "ridge_values": [0.0001, 0.001, 0.01, 0.1, 1.0, 10.0, 100.0], "validation_selects_ridge": true, "worker_processes": 4, "seed": 20260724}, "neural_ranker_dataset": {"expressions": 30000, "groups": 60000, "candidates": 299645, "valid_candidates": 298211, "failed_candidates": 1434, "empty_groups": 23132, "replay_mismatches": 0, "split_groups": {"train": 21568, "validation": 18168, "test_iid": 18050, "test_ood": 2214}}, "neural_ranker_metrics": {"test_iid": {"denominator_groups": 18050, "evaluable": 10752, "unevaluable": 7298, "neural_exact_best": 8349, "neural_exact_best_rate": 0.776506696429, "neural_mean_regret": 2.94540550595, "neural_total_regret": 31669, "neural_speedup": 15.0162, "estimated_eml_tree_exact_best": 8796, "estimated_eml_tree_mean_regret": 2.69280133929, "ast_dag_exact_best": 8661, "ast_dag_mean_regret": 2.18740699405, "random_exact_best": 2158, "random_mean_regret": 53.4677269345}, "test_ood": {"denominator_groups": 2214, "evaluable": 2034, "unevaluable": 180, "neural_exact_best": 1499, "neural_exact_best_rate": 0.736971484759, "neural_mean_regret": 3.09636184857, "neural_total_regret": 6298, "neural_speedup": 5.15466, "estimated_eml_tree_exact_best": 1555, "estimated_eml_tree_mean_regret": 4.5191740413, "ast_dag_exact_best": 1565, "ast_dag_mean_regret": 2.60078662734, "random_exact_best": 1078, "random_mean_regret": 10.994100295}}, "neural_ranker_runtime": {"candidate_scoring_seconds": 7102.153, "replay_seconds": 10721.489, "model_fit_eval_seconds": 24.532, "peak_rss_bytes": 1700249600, "workers": 4}, "production_export": {"dataset_id": "geml-250k-goal5", "expressions": 250000, "graphs": 1250000, "hierarchies": 250000, "batches": 2500, "validation_failures": 0, "reconstruction_failures": 0, "runtime": null, "peak_memory": null} }, "phase_b": { "source": "Phase_B_GPU_results/phase-b-20260731-anonymous/FINAL_VALIDATION.json", "runtime": {"snapshot": "ANON-PHASE-B-SNAPSHOT", "hardware": "4x NVIDIA H100 80GB HBM3", "python": "3.12.12", "torch": "2.7.1+cu128", "worktree_clean": true}, "validation_baseline": {"pytest": "2833 passed", "ruff_check": "passed", "ruff_format": "passed"}, "goal6": {"status": "complete_source_refresh_only", "common_harness": {"max_epochs": 30, "precision_config": "float32", "production_intended_precision": "bfloat16", "max_nodes_per_batch": 4096, "max_edges_per_batch": 8192, "max_examples_per_batch": 64, "gradient_accumulation_steps": 1, "optimizer": "adamw", "learning_rate": 0.001, "weight_decay": 0.01, "gradient_clip_norm": 1.0, "total_optimizer_steps": 20000, "parameter_tolerance": 0.15, "flop_tolerance": 0.25}, "architecture": {"graph_backbone": "compact_graph_encoder", "width": null, "permitted_widths": [64, 96], "virtual_node": null, "status": "pending_preflight_freeze"}, "cells": [{"seed": 20260726, "validation_loss": 2.391449174101672, "wall_seconds": 2883.607450513169, "checkpoint_sha256": "4d5d5a899b51f2d886fe1201cb73443c0776944d8f158591a8e9bdcfa1118507"}, {"seed": 20260727, "validation_loss": 0.5267228093910907, "wall_seconds": 2890.861278101802, "checkpoint_sha256": "69fcd7e336653fe14fad062c9144cd902d1357261ec904592109c69260f7f608"}, {"seed": 20260728, "validation_loss": 0.5386937659200733, "wall_seconds": 2888.853075893596, "checkpoint_sha256": "01d2385d3a33a3f1ed61c00e852a09e747050abdb6ccb00ddfaacb75760af751"}]}, "goal7": {"logical_cells": 18, "complete": 13, "invalid": 5, "raw_status_records": 26, "retained_failed_attempt_artifacts": 8, "attempt_failed_statuses": 5, "auth_errors": 0, "gpu_indices": [0, 1, 2, 3], "interrupted": false, "run_id": "5ad0f28779709c2c3efb5c6eb614a3c1f1017422dc6fde92b952bac1e42c81c0", "scheduler_state": "incomplete", "summary_sha256": "da792d92216ddfc9f7506f1fabf5085ff4196964ab3e604bc09aff7616e86a90"}, "retrieval": {"logical_cells": 15, "complete": 15, "auth_errors": 0, "gpu_indices": [0, 1, 2, 3], "interrupted": false, "summary_sha256": "464fb5bbd353239007408a71dba4a01cc0c26bd42ba66478a3a4603e7edba4ee"}, "goal8": {"status": "complete_runs_null_scientific_result", "architecture_config": {"shared_encoder": "goal6_compact_gnn", "embedding_dimension": null, "goal_conditioned": true, "composition": ["current_embedding", "goal_embedding", "current_minus_goal", "current_times_goal"], "hidden_dimension": 32, "ordinal_thresholds": 16, "value_head_parameter_cap": 20000, "maximum_total_trainable_parameters": 1000000, "regression": "one_sided_huber_for_right_censored_targets", "huber_delta": 1.0, "regression_weight": 1.0, "ordinal": "binary_cross_entropy_over_remaining_distance_thresholds", "ordinal_weight": 0.25}, "preparation": {"training_expressions": 35811, "train": 25342, "validation": 2540, "evaluation": 22570, "excluded_family": 21435, "excluded_length": 72431, "held_out_families": ["exp_log"], "maximum_witness_length": 4, "vocabulary_sha256": "88d709a7db1cd5e7982d32d6030633fafca605423791d0b5e7039afceadc3626"}, "training": {"runs": 3, "optimizer_steps": 1500, "epochs_completed": 2, "parameter_count": 452820, "seeds": [20260726, 20260727, 20260728], "selected_seed": 20260728}, "cells": [{"seed": 20260726, "validation_mae": 0.8471730067035345, "test_iid_mae": 1.9161517538686852, "test_iid_spearman": -0.014220334119549668, "test_ood_mae": 1.9076318031805681, "test_ood_spearman": null, "test_iid_n": 11218, "test_ood_n": 11352, "wall_seconds": 4560.555865172297, "best_epoch": 0, "checkpoint_sha256": "b75092f168f2d34ba49153bff87bb276045634e24c5fcf4f98ecabbfd94a53b1", "cell_sha256": "8231f8b972c0cb86ce0d8d8b6bf581a2e5ad722237963d56b724a679f2d52131"}, {"seed": 20260727, "validation_mae": 0.8375518081695076, "test_iid_mae": 1.8886932196262622, "test_iid_spearman": 0.002107999223893429, "test_ood_mae": 1.8804192692566455, "test_ood_spearman": null, "test_iid_n": 11218, "test_ood_n": 11352, "wall_seconds": 4512.926948711276, "best_epoch": 1, "checkpoint_sha256": "8dbd794ce901df021d8d3a944b7772348e0d16cb32f31dc7e4b260764de50d58", "cell_sha256": "1c25f4d2f224d57878306347b0425c386a2e82fe67681b4a0c40d2508a8024c5"}, {"seed": 20260728, "validation_mae": 0.8295553917021263, "test_iid_mae": 1.8684729886344036, "test_iid_spearman": -0.008904859511000642, "test_ood_mae": 1.8600275619001436, "test_ood_spearman": null, "test_iid_n": 11218, "test_ood_n": 11352, "wall_seconds": 4490.525827109814, "best_epoch": 0, "checkpoint_sha256": "a0aa9e78a5b11624e4fc58475d77000aef15ddd2dd1fb8b2816692faec34e9d9", "cell_sha256": "7efe85dfc8e5a750b00b26edb677f48b77db7f3b8256418b0057e4a96ecbb900"}], "selected_checkpoint_sha256": "a0aa9e78a5b11624e4fc58475d77000aef15ddd2dd1fb8b2816692faec34e9d9", "selected_source_encoder_sha256": "69fcd7e336653fe14fad062c9144cd902d1357261ec904592109c69260f7f608", "scientific_review": "null_or_collapsed_value_head"} }, "phase_b_integrity": {"goal6": {"seed_20260726": {"checkpoint_receipt_sha256": "12d7fbe542a42e3461fb035cb6ef9d89de003aed3021f18b09d1f6ceb71a2151", "row_sha256": "3b7868c7d905a7264e85c1d627de4a1c61954fecb1cd339b9348deb058033680"}, "seed_20260727": {"checkpoint_receipt_sha256": "05cff2a605f9dc4304ca63971f4f70f203ddfd0f1e7b47b513d2361dc9f95f9a", "row_sha256": "a1d5ac70fc33f6c485150b28432cca050cd3340580e421bcb8d2988b36ad56ae"}, "seed_20260728": {"checkpoint_receipt_sha256": "95bddb47d9ff824c42a77462944bfcf8dbfc665aac8fe865cdc71da1cd7441d9", "row_sha256": "64dd65a319263ca6e5fc7a3a108844dc9c761884db328c5e79ee8f6b7c1420e0"}}, "goal7_summary_sha256": "da792d92216ddfc9f7506f1fabf5085ff4196964ab3e604bc09aff7616e86a90", "retrieval_summary_sha256": "464fb5bbd353239007408a71dba4a01cc0c26bd42ba66478a3a4603e7edba4ee", "goal8_preparation_sha256": "7343c73ff73a9c37ca9f27474e293c789068d7711270f3727d6a2ac64e43bdaf", "goal8_step_dataset_digest": "ab15c6084cba1695b691cb25662881e68b3465e67c4536cd50b245759c3cb2bd", "goal8_step_manifest_sha256": "f79bf6510877399bb8775d1381b4ee120066d3be0a46fcd9b5d0974a892bacb0", "goal8_source_reconciliation_sha256": "5ba3a77c47f27f2d98cd4a2db713be13d7b619b6024d908cdd66ea02475b9b01", "goal8_summary_sha256": "aaae208091d5ae9f8657a31bad645457d712db2b1e2912b78267138ff198d799"}, "phase_b_goal8_extensions": {"failed_epochs_total": 0, "optimizer_steps_per_cell": 1500, "epochs_completed_per_cell": 2, "parameter_count": 452820, "value_head_parameter_cap": 20000, "maximum_total_trainable_parameters": 1000000, "evaluation": {"test_iid": {"attempted_per_cell": 11218, "valid_per_cell": 11218, "failure_per_cell": 0, "unsupported_per_cell": 0}, "test_ood": {"attempted_per_cell": 11352, "valid_per_cell": 11352, "failure_per_cell": 0, "unsupported_per_cell": 0}}}, "phase_b_release": {"schema_version": "geml-phase-b-final-validation-v1", "artifact_authentication": "passed", "operational_completion": "complete", "scientific_interpretation": "goal8_value_head_is_a_retained_null_result_not_a_useful_ranker", "runtime_source_snapshot": "ANON-PHASE-B-SNAPSHOT", "goal7_expected_cells": 18, "goal7_scheduler_state": "incomplete", "goal8_status_counts": {"complete": 3}, "retrieval_status_counts": {"complete": 15}, "gpu_indices": [0, 1, 2, 3], "interrupted": false}, "goal8_parameter_caps": {"value_head_only_cap": 20000, "maximum_total_trainable_parameters": 1000000, "observed_total_trainable_parameters": 452820, "interpretation": "The 20,000 cap applies only to the value head; the observed full model total is compared with the separate 1,000,000 cap."}, "goals9_12": {"goal9": {"status": "not_authenticated", "available_metrics": null}, "goal10": {"status": "not_complete", "published_endpoint_issue": "8 asin/acos endpoint cells", "corpus_or_learning_rerun": null}, "goal11": {"status": "not_authenticated", "external_llm_rows": null}, "goal12": {"status": "not_authenticated", "release_checklist": "unresolved"}}, "warnings": ["Goal7 invalid cells are retained in the denominator.", "GPU bitwise reproducibility across hardware is not claimed; cuBLAS deterministic warnings remain in logs.", "Goal8 held-out value predictions are nearly constant; OOD Spearman is undefined and the result is null/collapsed.", "Goal5 learned motif MDL mean denominator is 75000, not 250000.", "Goal5 export runtime and peak memory are unavailable in the authenticated production manifest."] }