Update: 1,651 SFT pairs, 2,576 corpus records (ORC+OAC+CL), Qwopus judge
Browse files
README.md
CHANGED
|
@@ -51,13 +51,22 @@ OhioCustodyBERT is one of **4 models** in a multi-agent legal council architectu
|
|
| 51 |
| Ohio Supreme Court | 199 | Shearer, Felton, In re Murray, In re Bonfield |
|
| 52 |
| US Supreme Court | 121 | Santosky v. Kramer, Stanley v. Illinois, Troxel v. Granville |
|
| 53 |
|
| 54 |
-
### SFT Training Data (audited March
|
| 55 |
| Dataset | Records | Status |
|
| 56 |
|---|---|---|
|
| 57 |
| v19 curated pairs | 916 train / 76 valid | ✅ Audited (77% clean) |
|
| 58 |
| Opinion-generated pairs (glm-5:cloud) | 480 (469/469 opinions) | ✅ Complete (96.6% clean) |
|
| 59 |
| Phase 2 strategy pairs | 177 | ✅ Ready |
|
| 60 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 61 |
|
| 62 |
### Reference Materials
|
| 63 |
- **941 records (4.0 MB)**: Black's Law Dictionary (591), OH Task Force on Family Law, Custody Evaluator Toolkit, GAL Task Force, Parenting Time Guide
|
|
|
|
| 51 |
| Ohio Supreme Court | 199 | Shearer, Felton, In re Murray, In re Bonfield |
|
| 52 |
| US Supreme Court | 121 | Santosky v. Kramer, Stanley v. Illinois, Troxel v. Granville |
|
| 53 |
|
| 54 |
+
### SFT Training Data (audited March 12, 2026)
|
| 55 |
| Dataset | Records | Status |
|
| 56 |
|---|---|---|
|
| 57 |
| v19 curated pairs | 916 train / 76 valid | ✅ Audited (77% clean) |
|
| 58 |
| Opinion-generated pairs (glm-5:cloud) | 480 (469/469 opinions) | ✅ Complete (96.6% clean) |
|
| 59 |
| Phase 2 strategy pairs | 177 | ✅ Ready |
|
| 60 |
+
| OFR report pairs (glm-5:cloud) | 115 (117 reports) | ✅ Complete |
|
| 61 |
+
| **Packaged total** | **1,536 train / 115 valid** | **1,651 unique pairs** |
|
| 62 |
+
|
| 63 |
+
### Cloud-Scraped Corpus (March 12, 2026)
|
| 64 |
+
| Source | Records | Size |
|
| 65 |
+
|---|---|---|
|
| 66 |
+
| Ohio Revised Code (19 chapters) | 1,034 sections | 3.0 MB |
|
| 67 |
+
| Ohio Administrative Code (19 divisions) | 1,283 rules | 7.3 MB |
|
| 68 |
+
| CourtListener opinions (OH App, SC, SCOTUS) | 259 opinions | 14 MB |
|
| 69 |
+
| **Total corpus** | **2,576 records** | **24.3 MB** |
|
| 70 |
|
| 71 |
### Reference Materials
|
| 72 |
- **941 records (4.0 MB)**: Black's Law Dictionary (591), OH Task Force on Family Law, Custody Evaluator Toolkit, GAL Task Force, Parenting Time Guide
|