{ "strategy": "document-level hash bucketing, seed chosen for label coverage", "ratios": [ 0.7, 0.15, 0.15 ], "seed": 42, "documents": { "train": 55289, "validation": 11705, "test": 11901 }, "examples": { "train": 61483, "validation": 13151, "test": 13481 }, "label_distribution": { "train": { "OTHER": 11022, "TERM_AND_DURATION": 2108, "RENEWAL": 124, "GOVERNING_LAW": 3195, "EXCLUSIVITY": 525, "NON_SOLICITATION": 89, "ASSIGNMENT": 2005, "PAYMENT": 3983, "SCOPE_OF_WORK": 895, "LICENSE_GRANT": 561, "SURVIVAL": 1614, "WARRANTY": 572, "INSURANCE": 1388, "DISPUTE_RESOLUTION": 3744, "CHANGE_OF_CONTROL": 487, "AUDIT_RIGHTS": 737, "LIABILITY": 512, "TERMINATION": 1640, "NON_COMPETE": 241, "INTELLECTUAL_PROPERTY": 679, "PUBLICITY": 672, "WAIVER": 1330, "COMPLIANCE": 2306, "COUNTERPARTS": 2269, "INDEMNIFICATION": 1086, "REPRESENTATIONS": 4611, "WORKING_ARRANGEMENT": 572, "ENTIRE_AGREEMENT": 2655, "NOTICES": 2292, "SEVERABILITY": 1988, "TAXES": 1847, "EXPENSES": 1206, "AMENDMENT": 1592, "CONFIDENTIALITY": 936 }, "validation": { "OTHER": 2231, "ASSIGNMENT": 456, "TERM_AND_DURATION": 405, "GOVERNING_LAW": 696, "PAYMENT": 871, "EXCLUSIVITY": 109, "CHANGE_OF_CONTROL": 114, "LICENSE_GRANT": 132, "SURVIVAL": 358, "DISPUTE_RESOLUTION": 844, "NON_COMPETE": 61, "NON_SOLICITATION": 23, "TERMINATION": 353, "SCOPE_OF_WORK": 185, "INTELLECTUAL_PROPERTY": 157, "AUDIT_RIGHTS": 164, "LIABILITY": 170, "WARRANTY": 133, "INSURANCE": 324, "RENEWAL": 23, "PUBLICITY": 130, "COUNTERPARTS": 481, "COMPLIANCE": 510, "EXPENSES": 262, "REPRESENTATIONS": 959, "ENTIRE_AGREEMENT": 556, "WAIVER": 260, "NOTICES": 458, "TAXES": 418, "CONFIDENTIALITY": 185, "WORKING_ARRANGEMENT": 154, "INDEMNIFICATION": 226, "AMENDMENT": 343, "SEVERABILITY": 400 }, "test": { "OTHER": 2296, "TERM_AND_DURATION": 493, "GOVERNING_LAW": 714, "LIABILITY": 128, "INSURANCE": 244, "RENEWAL": 41, "TERMINATION": 374, "ASSIGNMENT": 471, "PAYMENT": 876, "AUDIT_RIGHTS": 161, "SCOPE_OF_WORK": 205, "LICENSE_GRANT": 213, "SURVIVAL": 328, "NON_COMPETE": 47, "EXCLUSIVITY": 118, "CHANGE_OF_CONTROL": 113, "INTELLECTUAL_PROPERTY": 167, "DISPUTE_RESOLUTION": 871, "PUBLICITY": 145, "WARRANTY": 130, "NON_SOLICITATION": 15, "NOTICES": 480, "REPRESENTATIONS": 974, "COUNTERPARTS": 488, "CONFIDENTIALITY": 177, "ENTIRE_AGREEMENT": 576, "AMENDMENT": 342, "WAIVER": 335, "WORKING_ARRANGEMENT": 110, "INDEMNIFICATION": 235, "COMPLIANCE": 509, "TAXES": 400, "EXPENSES": 289, "SEVERABILITY": 416 } }, "ungrouped_datasets": [ "LEDGAR" ], "leakage_check": "PASS - no document id appears in more than one split", "warnings": [ "No document grouping was available for: LEDGAR. Their examples were split individually, so metrics on those datasets may be optimistic relative to a true document-level split." ] }