Add 841-dev naked vs hybrid comparison with GLiNER and GLiNER 2.5

README.md

@@ -55,17 +55,23 @@ Fresh XLM-R, seven epochs. 133 epoch/threshold combinations. Epoch 5 at 0.95 was
 | 6 | 328 | 20 | 147 | 14 |
 | 7 | 334 | 16 | 148 | 15 |
 
-## 841-dev scores
+## Compared with other systems
 
-Union with unchanged rules. Character scores are on gold vs masked characters.
+Same 841-dev split, threshold **0.95**. **Naked** is the transformer alone. **∪ regex** is that model unioned with the frozen rules (the NERGAL recipe). Character scores are gold vs masked characters.
 
-| System | Whole spans /354 | Residual passages | False chars | Char precision | Char recall |
-|---|---:|---:|---:|---:|---:|
-| Regex only | 245 | 73 | 133 | 97.36% | 81.01% |
-| Historical GLiNER email12 ∪ regex | 289 | 51 | 143 | 97.54% | 93.48% |
-| **NERGAL (this snapshot ∪ regex)** | **323** | **25** | **133** | **97.76%** | **95.95%** |
+| System | Mode | Whole /354 | Residual | False chars | Char P | Char R |
+|---|---|---:|---:|---:|---:|---:|
+| Regex (`scrub_pii`) | rules | 245 | 73 | 133 | 97.36% | 81.01% |
+| GLiNER 2.5-multi zero-shot | naked | 81 | 188 | 970 | 56.98% | 21.22% |
+| GLiNER 2.5-multi zero-shot | ∪ regex | 262 | 63 | 1,103 | 82.36% | 85.02% |
+| Historical GLiNER email12 | naked | 249 | 70 | 10 | 99.81% | 85.39% |
+| Historical GLiNER email12 | ∪ regex | 289 | 51 | 143 | 97.54% | 93.48% |
+| XLM-R epoch 5 | naked | 298 | 42 | 57 | 98.96% | 89.38% |
+| **NERGAL (this snapshot)** | **∪ regex** | **323** | **25** | **133** | **97.76%** | **95.95%** |
 
-144/169 phone, 179/185 other PII. Exact-span precision 87.50%, recall 88.98%, F1 88.24%.
+Zero-shot [GLiNER 2.5](https://huggingface.co/fastino/gliner2.5-multi-v1) is not competitive here, especially on non-phone PII (7/185 whole vs 160 naked / 179 union for this snapshot). Fine-tuned historical GLiNER is the precise naked baseline (10 false characters) and still trails XLM-R on coverage. NERGAL is XLM-R epoch 5 plus the regex: 144/169 phone, 179/185 other PII. Exact-span precision 87.50%, recall 88.98%, F1 88.24%. The 133 false characters are the glued-email regex error; this seed does not add to that floor.
+
+Trained GLiNER, HerBERT-large, and XLM-R-large were also compared on this split (plot above). GLiNER’s 0.50 coverage lead is the precision collapse in that figure; no GLiNER or HerBERT operating point passed the content-preservation gate.
 
 ## Extra seeds