| { | |
| "experiment": "R-MATH-INK-06-FORMULA-CONTEXT-CONTRACT-001", | |
| "generated_at": "2026-07-23T20:58:38.398037+00:00", | |
| "split_contract": "CROHME trainData writer-validation mode selection; testDataGT one-shot", | |
| "context_modes": [ | |
| "formula", | |
| "missing", | |
| "local_full", | |
| "height_only" | |
| ], | |
| "validation_sweep": [ | |
| { | |
| "mode": "formula", | |
| "all_supported": { | |
| "samples": 3241, | |
| "exact_top1": 0.5637149028077754, | |
| "visual_family_top1": 0.6457883369330454 | |
| }, | |
| "behavior_targets": { | |
| "samples": 467, | |
| "exact_top1": 0.284796573875803, | |
| "visual_family_top1": 0.6552462526766595 | |
| } | |
| }, | |
| { | |
| "mode": "missing", | |
| "all_supported": { | |
| "samples": 3241, | |
| "exact_top1": 0.6078370873187288, | |
| "visual_family_top1": 0.7655044739278001 | |
| }, | |
| "behavior_targets": { | |
| "samples": 467, | |
| "exact_top1": 0.18629550321199143, | |
| "visual_family_top1": 0.44325481798715205 | |
| } | |
| }, | |
| { | |
| "mode": "local_full", | |
| "all_supported": { | |
| "samples": 3241, | |
| "exact_top1": 0.5387226164763962, | |
| "visual_family_top1": 0.6525763653193459 | |
| }, | |
| "behavior_targets": { | |
| "samples": 467, | |
| "exact_top1": 0.08137044967880086, | |
| "visual_family_top1": 0.5396145610278372 | |
| } | |
| }, | |
| { | |
| "mode": "height_only", | |
| "all_supported": { | |
| "samples": 3241, | |
| "exact_top1": 0.5029311940759025, | |
| "visual_family_top1": 0.5766738660907127 | |
| }, | |
| "behavior_targets": { | |
| "samples": 467, | |
| "exact_top1": 0.042826552462526764, | |
| "visual_family_top1": 0.41541755888650966 | |
| } | |
| } | |
| ], | |
| "selected_mode": "formula", | |
| "selected_validation": { | |
| "mode": "formula", | |
| "all_supported": { | |
| "samples": 3241, | |
| "exact_top1": 0.5637149028077754, | |
| "visual_family_top1": 0.6457883369330454 | |
| }, | |
| "behavior_targets": { | |
| "samples": 467, | |
| "exact_top1": 0.284796573875803, | |
| "visual_family_top1": 0.6552462526766595 | |
| } | |
| }, | |
| "validation_branch_oracle": { | |
| "all_supported": { | |
| "samples": 3241, | |
| "exact_oracle": 0.7682813946312866, | |
| "visual_family_oracle": 0.8805924097500771 | |
| }, | |
| "behavior_targets": { | |
| "samples": 467, | |
| "exact_oracle": 0.37687366167023556, | |
| "visual_family_oracle": 0.7516059957173448 | |
| } | |
| }, | |
| "official_test": { | |
| "all_supported": { | |
| "samples": 5921, | |
| "exact_top1": 0.6302989359905421, | |
| "visual_family_top1": 0.6818105049822665 | |
| }, | |
| "behavior_targets": { | |
| "samples": 635, | |
| "exact_top1": 0.3543307086614173, | |
| "visual_family_top1": 0.5574803149606299 | |
| } | |
| }, | |
| "track": "R_noncommercial_only", | |
| "product_validation": false, | |
| "distillation_allowed": false | |
| } | |