Thorsu commited on
Commit
e456c13
·
verified ·
1 Parent(s): 8772c0c

Record v0.4.1 split materialization

Browse files
Files changed (1) hide show
  1. source_registry.json +31 -4
source_registry.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
  "dataset": "ouroboroscollective/evidence-bound-css",
3
- "release": "0.4.0",
4
  "release_date": "2026-10-05",
5
  "canonical_unique_records": 601,
6
  "configs": {
@@ -75,7 +75,12 @@
75
  "test": 20
76
  },
77
  "split_policy": "family_disjoint",
78
- "benchmark_release": "0.4.0"
 
 
 
 
 
79
  },
80
  "benchmark_hard_negatives_de": {
81
  "rows": 120,
@@ -86,7 +91,12 @@
86
  "test": 20
87
  },
88
  "split_policy": "family_disjoint",
89
- "benchmark_release": "0.4.0"
 
 
 
 
 
90
  },
91
  "benchmark_preferences_de": {
92
  "rows": 120,
@@ -97,7 +107,12 @@
97
  "test": 20
98
  },
99
  "split_policy": "family_disjoint",
100
- "benchmark_release": "0.4.0"
 
 
 
 
 
101
  }
102
  },
103
  "source_inventory": {
@@ -217,6 +232,18 @@
217
  "all_after_fixes_passed": true
218
  },
219
  "truth_boundary": "Replay covers the 20 family-disjoint repair test rows only; it does not establish contamination-freedom against external model training corpora."
 
 
 
 
 
 
 
 
 
 
 
 
220
  }
221
  },
222
  "split_policy": "Single-source knowledge corpus uses train only. Future evals must be source- and family-disjoint and near-deduplicated.",
 
1
  {
2
  "dataset": "ouroboroscollective/evidence-bound-css",
3
+ "release": "0.4.1",
4
  "release_date": "2026-10-05",
5
  "canonical_unique_records": 601,
6
  "configs": {
 
75
  "test": 20
76
  },
77
  "split_policy": "family_disjoint",
78
+ "benchmark_release": "0.4.0",
79
+ "split_files": {
80
+ "train": "data/benchmark_repairs_de/train.jsonl",
81
+ "validation": "data/benchmark_repairs_de/validation.jsonl",
82
+ "test": "data/benchmark_repairs_de/test.jsonl"
83
+ }
84
  },
85
  "benchmark_hard_negatives_de": {
86
  "rows": 120,
 
91
  "test": 20
92
  },
93
  "split_policy": "family_disjoint",
94
+ "benchmark_release": "0.4.0",
95
+ "split_files": {
96
+ "train": "data/benchmark_hard_negatives_de/train.jsonl",
97
+ "validation": "data/benchmark_hard_negatives_de/validation.jsonl",
98
+ "test": "data/benchmark_hard_negatives_de/test.jsonl"
99
+ }
100
  },
101
  "benchmark_preferences_de": {
102
  "rows": 120,
 
107
  "test": 20
108
  },
109
  "split_policy": "family_disjoint",
110
+ "benchmark_release": "0.4.0",
111
+ "split_files": {
112
+ "train": "data/benchmark_preferences_de/train.jsonl",
113
+ "validation": "data/benchmark_preferences_de/validation.jsonl",
114
+ "test": "data/benchmark_preferences_de/test.jsonl"
115
+ }
116
  }
117
  },
118
  "source_inventory": {
 
232
  "all_after_fixes_passed": true
233
  },
234
  "truth_boundary": "Replay covers the 20 family-disjoint repair test rows only; it does not establish contamination-freedom against external model training corpora."
235
+ },
236
+ "v041_split_materialization": {
237
+ "status": "fixed",
238
+ "physical_split_files": true,
239
+ "rows_per_benchmark_config": {
240
+ "train": 80,
241
+ "validation": 20,
242
+ "test": 20
243
+ },
244
+ "family_assignment_unchanged": true,
245
+ "leakage_checks_unchanged": true,
246
+ "test_replay_unchanged": "20/20 PASS"
247
  }
248
  },
249
  "split_policy": "Single-source knowledge corpus uses train only. Future evals must be source- and family-disjoint and near-deduplicated.",