# scanpatch - ru / pii (1500 rows, 8708 spans, 171 negatives)

Main metric: **missed** - gold spans no predicted character touched. Score threshold 0.5; spans without a score always count. `hidden` - gold spans every character of which is covered: touching one character counts as detection, not as hiding. `hidden` is counted over normalized predictions - a prediction is stretched to whole words before it is compared, so it is what a masker that repeats the same normalization would hide; a masker that masks the raw offsets of the model hides no more than this. Char P / R / F1 are reference only. Trivial baseline, mask everything: 0 missed, 100% hidden, char P 0.176 (82.4% of the text over-masked). `rows touched` / `chars masked` count rows the source left without annotations where a mask touched something, and the characters masked in them; that absence of annotations is not evidence that those rows hold nothing sensitive, so this is not a false-alarm rate. The exact CSVs keep the technical column names `fp_rows` and `fp_chars`.

| model | missed | missed % [95% CI] | hidden | char F1 [95% CI] | P | R | entity exact | entity overlap | typed F1 | rows touched | chars masked | dropped spans | ms/row | params | train |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| nym-base | **776** | 8.9% [8.0%, 9.8%] | 70.7% | 0.856 [0.848, 0.863] | 0.885 | 0.828 | 0.579 | 0.904 | 0.398 | 35/171 | 380 | 0 | 4 | 308M |  |
| nym-small | **904** | 10.4% [9.4%, 11.3%] | 71.3% | 0.846 [0.838, 0.854] | 0.884 | 0.811 | 0.570 | 0.895 | 0.381 | 30/171 | 348 | 0 | 73 | - |  |
| pplx | **1084** | 12.4% [11.2%, 13.7%] | 85.5% | 0.862 [0.854, 0.871] | 0.875 | 0.850 | 0.627 | 0.892 | 0.815 | 28/171 | 337 | 0 | 59 | 596M |  |
| mmbert32k | **1413** | 16.2% [15.2%, 17.3%] | 50.4% | 0.758 [0.749, 0.767] | 0.903 | 0.653 | 0.334 | 0.845 | 0.478 | 41/171 | 351 | 0 | 6 | 308M |  |
| nuner-zero | **1529** | 17.6% [16.3%, 19.0%] | 76.2% | 0.803 [0.793, 0.813] | 0.812 | 0.794 | 0.572 | 0.831 | 0.769 | 125/171 | 2031 | 0 | 14 | 449M |  |
| bardsai-eu | **1591** | 18.3% [16.8%, 19.7%] | 72.6% | 0.801 [0.790, 0.813] | 0.886 | 0.732 | 0.533 | 0.858 | 0.264 | 30/171 | 329 | 0 | 336 | - |  |
| gliner2-fastino | **1638** | 18.8% [17.6%, 20.1%] | 74.2% | 0.781 [0.769, 0.793] | 0.858 | 0.717 | 0.686 | 0.838 | 0.751 | 73/171 | 936 | 0 | 8 | 307M |  |
| apararti | **1642** | 18.9% [17.7%, 20.1%] | 74.4% | 0.817 [0.808, 0.827] | 0.831 | 0.805 | 0.450 | 0.845 | 0.681 | 30/171 | 607 | 0 | 63 | 1.4B |  |
| ru-legal-ner+cpu | **1724** | 19.8% [18.5%, 21.1%] | 62.2% | 0.743 [0.730, 0.756] | 0.842 | 0.664 | 0.347 | 0.840 | 0.585 | 33/171 | 444 | 0 | 11 | 29M |  |
| ru-legal-ner | **1724** | 19.8% [18.5%, 21.1%] | 62.2% | 0.743 [0.730, 0.756] | 0.842 | 0.664 | 0.347 | 0.840 | 0.585 | 33/171 | 444 | 0 | 3 | 29M |  |
| pii-shield-onnx | **1964** | 22.6% [21.3%, 23.8%] | 58.2% | 0.789 [0.779, 0.799] | 0.911 | 0.696 | 0.389 | 0.835 | 0.500 | 30/171 | 408 | 0 | 132 | - |  |
| opf-ru | **2044** | 23.5% [22.2%, 24.6%] | 54.7% | 0.712 [0.700, 0.724] | 0.880 | 0.598 | 0.397 | 0.820 | 0.583 | 68/171 | 713 | 0 | 181 | 1.4B |  |
| openmed-multilingual | **2071** | 23.8% [22.4%, 25.2%] | 51.4% | 0.775 [0.766, 0.785] | 0.882 | 0.692 | 0.461 | 0.812 | 0.699 | 57/171 | 613 | 0 | 145 | 1.4B |  |
| gliner-nvidia | **2172** | 24.9% [23.5%, 26.4%] | 57.3% | 0.718 [0.706, 0.729] | 0.816 | 0.641 | 0.552 | 0.791 | 0.686 | 97/171 | 1113 | 0 | 10 | 445M |  |
| openai-base | **2230** | 25.6% [24.0%, 27.2%] | 70.2% | 0.775 [0.764, 0.787] | 0.822 | 0.734 | 0.448 | 0.818 | 0.698 | 19/171 | 340 | 0 | 273 | 1.4B |  |
| openmed-nemotron | **2364** | 27.1% [25.8%, 28.5%] | 48.9% | 0.691 [0.680, 0.702] | 0.741 | 0.648 | 0.390 | 0.734 | 0.600 | 76/171 | 992 | 0 | 147 | 1.4B |  |
| gliner-urchade | **2389** | 27.4% [26.1%, 28.8%] | 66.7% | 0.696 [0.683, 0.709] | 0.751 | 0.649 | 0.517 | 0.744 | 0.664 | 120/171 | 1845 | 0 | 5 | 289M |  |
| gliner2-fastino-ru | **2673** | 30.7% [29.1%, 32.2%] | 61.5% | 0.719 [0.707, 0.732] | 0.861 | 0.618 | 0.567 | 0.778 | 0.699 | 57/171 | 844 | 0 | 5 | 307M |  |
| opf-kz-ru | **2723** | 31.3% [29.6%, 32.9%] | 59.3% | 0.760 [0.749, 0.771] | 0.845 | 0.690 | 0.385 | 0.769 | 0.667 | 33/171 | 498 | 0 | 55 | 1.4B |  |
| ru-pii-ner | **2797** | 32.1% [30.1%, 34.2%] | 65.9% | 0.707 [0.690, 0.724] | 0.872 | 0.595 | 0.524 | 0.780 | 0.439 | 34/171 | 348 | 0 | 62 | 358M |  |
| traciora | **2830** | 32.5% [30.6%, 34.3%] | 59.2% | 0.685 [0.670, 0.702] | 0.826 | 0.585 | 0.376 | 0.773 | 0.626 | 50/171 | 808 | 0 | 246 | 1.4B |  |
| gliner-pii-edge | **2847** | 32.7% [31.2%, 34.2%] | 57.0% | 0.666 [0.654, 0.678] | 0.767 | 0.588 | 0.480 | 0.714 | 0.468 | 125/171 | 2042 | 0 | 87 | 45M |  |
| fef2-secret-ru | **2851** | 32.7% [30.8%, 34.6%] | 59.5% | 0.680 [0.663, 0.699] | 0.916 | 0.541 | 0.492 | 0.780 | 0.537 | 32/171 | 299 | 0 | 3 | 177M |  |
| opf-ru-v2 | **2857** | 32.8% [31.0%, 34.6%] | 56.5% | 0.715 [0.701, 0.729] | 0.884 | 0.601 | 0.386 | 0.771 | 0.653 | 16/171 | 269 | 0 | 466 | 1.4B |  |
| gliner2-hivetrace-omni | **2901** | 33.3% [31.7%, 35.1%] | 59.1% | 0.695 [0.681, 0.709] | 0.864 | 0.582 | 0.612 | 0.760 | 0.675 | 53/171 | 581 | 0 | 9 | 307M |  |
| gliner-nvidia-ru | **2909** | 33.4% [32.0%, 34.9%] | 49.5% | 0.682 [0.671, 0.694] | 0.880 | 0.557 | 0.538 | 0.764 | 0.613 | 48/171 | 429 | 0 | 46 | 445M |  |
| kalyan-ettin | **2933** | 33.7% [32.0%, 35.4%] | 42.5% | 0.698 [0.685, 0.710] | 0.891 | 0.574 | 0.390 | 0.757 | 0.638 | 48/171 | 467 | 0 | 14 | 68M |  |
| stanza-ru | **2972** | 34.1% [32.3%, 36.1%] | 57.8% | 0.624 [0.605, 0.643] | 0.818 | 0.505 | 0.442 | 0.734 | 0.378 | 73/171 | 952 | 0 | 42 | - |  |
| gliner25-fastino | **3036** | 34.9% [33.1%, 36.6%] | 61.0% | 0.723 [0.709, 0.736] | 0.874 | 0.616 | 0.554 | 0.755 | 0.607 | 57/171 | 784 | 0 | 3 | 287M |  |
| gliner2-hivetrace-omni-ru | **3286** | 37.7% [36.0%, 39.5%] | 54.1% | 0.664 [0.649, 0.677] | 0.868 | 0.537 | 0.559 | 0.734 | 0.647 | 54/171 | 558 | 0 | 11 | 307M |  |
| davlan-xlmr | **3289** | 37.8% [35.8%, 39.9%] | 58.2% | 0.636 [0.616, 0.655] | 0.913 | 0.488 | 0.461 | 0.746 | 0.440 | 31/171 | 291 | 0 | 4 | 277M |  |
| davlan-mbert | **3439** | 39.5% [37.4%, 41.6%] | 55.8% | 0.622 [0.602, 0.642] | 0.917 | 0.471 | 0.471 | 0.733 | 0.443 | 29/171 | 292 | 0 | 4 | 177M |  |
| gliner-multi-v21 | **3516** | 40.4% [38.8%, 42.0%] | 53.5% | 0.600 [0.585, 0.615] | 0.806 | 0.478 | 0.438 | 0.690 | 0.550 | 85/171 | 892 | 0 | 4 | 289M |  |
| gliner2-hivetrace-uni | **3598** | 41.3% [39.8%, 42.9%] | 54.2% | 0.632 [0.617, 0.646] | 0.874 | 0.495 | 0.529 | 0.708 | 0.551 | 42/171 | 369 | 0 | 18 | 147M |  |
| ner-ru-gherman | **3622** | 41.6% [39.6%, 43.6%] | 36.2% | 0.549 [0.529, 0.569] | 0.920 | 0.391 | 0.443 | 0.715 | 0.100 | 22/171 | 133 | 0 | 4 | 177M |  |
| spacy-ru-lg | **3706** | 42.6% [40.5%, 44.7%] | 50.9% | 0.589 [0.568, 0.609] | 0.891 | 0.439 | 0.435 | 0.704 | 0.396 | 43/171 | 497 | 0 | 5 | - |  |
| gliner25-fastino-ru | **3792** | 43.5% [41.9%, 45.4%] | 52.8% | 0.660 [0.645, 0.675] | 0.888 | 0.526 | 0.485 | 0.699 | 0.567 | 45/171 | 563 | 0 | 4 | 287M |  |
| gliner2-large | **3922** | 45.0% [43.3%, 46.9%] | 49.0% | 0.576 [0.561, 0.590] | 0.720 | 0.480 | 0.466 | 0.637 | 0.503 | 118/171 | 1878 | 0 | 23 | 486M |  |
| natasha | **3930** | 45.1% [43.0%, 47.2%] | 49.3% | 0.566 [0.545, 0.588] | 0.901 | 0.412 | 0.426 | 0.688 | 0.409 | 34/171 | 405 | 0 | 2 | - |  |
| spacy-alrosait | **4043** | 46.4% [44.3%, 48.6%] | 53.2% | 0.581 [0.560, 0.603] | 0.967 | 0.415 | 0.429 | 0.694 | 0.194 | 5/171 | 54 | 0 | 5 | - | train |
| gliner-urchade-ru | **4164** | 47.8% [46.0%, 49.5%] | 47.0% | 0.614 [0.601, 0.628] | 0.744 | 0.522 | 0.470 | 0.625 | 0.567 | 72/171 | 1444 | 0 | 10 | 289M |  |
| ner-ru-yqelz | **4176** | 48.0% [46.0%, 49.9%] | 44.9% | 0.531 [0.511, 0.551] | 0.803 | 0.396 | 0.350 | 0.643 | 0.352 | 73/171 | 1079 | 0 | 5 | 559M |  |
| gliner-stream-pii | **4213** | 48.4% [46.9%, 50.1%] | 38.8% | 0.539 [0.526, 0.553] | 0.855 | 0.394 | 0.368 | 0.649 | 0.480 | 140/171 | 1030 | 0 | 11 | 677M |  |
| gliner-multi-v21-ru | **4292** | 49.3% [47.4%, 51.0%] | 46.0% | 0.545 [0.530, 0.561] | 0.820 | 0.408 | 0.405 | 0.634 | 0.485 | 69/171 | 709 | 0 | 9 | 289M |  |
| gliner2-vladlinv | **4519** | 51.9% [49.5%, 54.5%] | 46.4% | 0.558 [0.534, 0.580] | 0.901 | 0.404 | 0.415 | 0.632 | 0.433 | 21/171 | 149 | 0 | 4 | 287M |  |
| gliner2-vladlinv-ru | **5514** | 63.3% [60.8%, 66.0%] | 35.4% | 0.473 [0.448, 0.499] | 0.901 | 0.321 | 0.355 | 0.524 | 0.412 | 19/171 | 147 | 0 | 4 | 287M |  |
| gravitee-small | **5578** | 64.1% [62.2%, 65.8%] | 24.8% | 0.460 [0.444, 0.476] | 0.671 | 0.350 | 0.207 | 0.493 | 0.407 | 72/171 | 1558 | 0 | 7 | 29M |  |
| gliner-pii-base | **5881** | 67.5% [65.8%, 69.2%] | 26.2% | 0.447 [0.431, 0.463] | 0.789 | 0.312 | 0.364 | 0.457 | 0.419 | 73/171 | 861 | 0 | 4 | 166M |  |
| gliner2-hivetrace-uni-ru | **6911** | 79.4% [78.0%, 80.7%] | 19.1% | 0.286 [0.270, 0.302] | 0.875 | 0.171 | 0.249 | 0.335 | 0.196 | 15/171 | 141 | 0 | 9 | 147M |  |
| rules-ru | **7442** | 85.5% [84.2%, 86.7%] | 11.7% | 0.364 [0.342, 0.387] | 0.904 | 0.228 | 0.199 | 0.250 | 0.362 | 3/171 | 46 | 0 | 0 | - |  |
| betterleaks | **8708** | 100.0% [100.0%, 100.0%] | 0.0% | 0.000 [0.000, 0.000] | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0/171 | 0 | 0 | 0 | - |  |
| gitleaks | **8708** | 100.0% [100.0%, 100.0%] | 0.0% | 0.000 [0.000, 0.000] | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0/171 | 0 | 0 | 0 | - |  |

Neighbours by rank a paired bootstrap cannot tell apart (95%, missed %) - pairs, not a transitive chain, so `a ≈ b` and `b ≈ c` do not make `a ≈ c`: mmbert32k ≈ nuner-zero; nuner-zero ≈ bardsai-eu; bardsai-eu ≈ gliner2-fastino; gliner2-fastino ≈ apararti; apararti ≈ ru-legal-ner+cpu; ru-legal-ner+cpu ≈ ru-legal-ner; pii-shield-onnx ≈ opf-ru; opf-ru ≈ openmed-multilingual; openmed-multilingual ≈ gliner-nvidia; gliner-nvidia ≈ openai-base; openai-base ≈ openmed-nemotron; openmed-nemotron ≈ gliner-urchade; gliner2-fastino-ru ≈ opf-kz-ru; opf-kz-ru ≈ ru-pii-ner; ru-pii-ner ≈ traciora; traciora ≈ gliner-pii-edge; gliner-pii-edge ≈ fef2-secret-ru; fef2-secret-ru ≈ opf-ru-v2; opf-ru-v2 ≈ gliner2-hivetrace-omni; gliner2-hivetrace-omni ≈ gliner-nvidia-ru; gliner-nvidia-ru ≈ kalyan-ettin; kalyan-ettin ≈ stanza-ru; stanza-ru ≈ gliner25-fastino; gliner2-hivetrace-omni-ru ≈ davlan-xlmr; davlan-mbert ≈ gliner-multi-v21; gliner-multi-v21 ≈ gliner2-hivetrace-uni; gliner2-hivetrace-uni ≈ ner-ru-gherman; spacy-ru-lg ≈ gliner25-fastino-ru; gliner25-fastino-ru ≈ gliner2-large; gliner2-large ≈ natasha; spacy-alrosait ≈ gliner-urchade-ru; gliner-urchade-ru ≈ ner-ru-yqelz; ner-ru-yqelz ≈ gliner-stream-pii; gliner-stream-pii ≈ gliner-multi-v21-ru; gliner-multi-v21-ru ≈ gliner2-vladlinv; gliner2-vladlinv-ru ≈ gravitee-small; betterleaks ≈ gitleaks

Legacy runs on this set: betterleaks, pii-shield-onnx, ru-legal-ner+cpu. Their meta carries no `protocol` and no `bench_sha256`, so the version of the text they were taken from is confirmed only by the row and character counts, not by a fingerprint.

## Missed by group

| model | PERSON | ADDRESS | CONTACT | ID | NET | ORG | OTHER |
|---|---|---|---|---|---|---|---|
| spans in gold | 3798 | 1637 | 588 | 655 | 1103 | 277 | 650 |
| nym-base | 27 (0.7%) | 116 (7.1%) | 15 (2.6%) | 117 (17.9%) | 388 (35.2%) | 76 (27.4%) | 37 (5.7%) |
| nym-small | 113 (3.0%) | 184 (11.2%) | 8 (1.4%) | 117 (17.9%) | 346 (31.4%) | 84 (30.3%) | 52 (8.0%) |
| pplx | 24 (0.6%) | 528 (32.3%) | 2 (0.3%) | 52 (7.9%) | 66 (6.0%) | 249 (89.9%) | 163 (25.1%) |
| mmbert32k | 620 (16.3%) | 520 (31.8%) | 6 (1.0%) | 52 (7.9%) | 46 (4.2%) | 130 (46.9%) | 39 (6.0%) |
| nuner-zero | 235 (6.2%) | 392 (23.9%) | 10 (1.7%) | 179 (27.3%) | 27 (2.4%) | 44 (15.9%) | 642 (98.8%) |
| bardsai-eu | 48 (1.3%) | 141 (8.6%) | 25 (4.3%) | 110 (16.8%) | 587 (53.2%) | 50 (18.1%) | 630 (96.9%) |
| gliner2-fastino | 179 (4.7%) | 187 (11.4%) | 5 (0.9%) | 131 (20.0%) | 451 (40.9%) | 41 (14.8%) | 644 (99.1%) |
| apararti | 749 (19.7%) | 528 (32.3%) | 5 (0.9%) | 30 (4.6%) | 33 (3.0%) | 172 (62.1%) | 125 (19.2%) |
| ru-legal-ner+cpu | 309 (8.1%) | 553 (33.8%) | 12 (2.0%) | 132 (20.2%) | 502 (45.5%) | 80 (28.9%) | 136 (20.9%) |
| ru-legal-ner | 309 (8.1%) | 553 (33.8%) | 12 (2.0%) | 132 (20.2%) | 502 (45.5%) | 80 (28.9%) | 136 (20.9%) |
| pii-shield-onnx | 834 (22.0%) | 479 (29.3%) | 9 (1.5%) | 179 (27.3%) | 115 (10.4%) | 223 (80.5%) | 125 (19.2%) |
| opf-ru | 424 (11.2%) | 485 (29.6%) | 15 (2.6%) | 155 (23.7%) | 397 (36.0%) | 184 (66.4%) | 384 (59.1%) |
| openmed-multilingual | 1172 (30.9%) | 455 (27.8%) | 17 (2.9%) | 77 (11.8%) | 38 (3.4%) | 178 (64.3%) | 134 (20.6%) |
| gliner-nvidia | 310 (8.2%) | 355 (21.7%) | 105 (17.9%) | 197 (30.1%) | 510 (46.2%) | 50 (18.1%) | 645 (99.2%) |
| openai-base | 831 (21.9%) | 678 (41.4%) | 15 (2.6%) | 95 (14.5%) | 203 (18.4%) | 210 (75.8%) | 198 (30.5%) |
| openmed-nemotron | 782 (20.6%) | 582 (35.6%) | 63 (10.7%) | 380 (58.0%) | 194 (17.6%) | 171 (61.7%) | 192 (29.5%) |
| gliner-urchade | 449 (11.8%) | 268 (16.4%) | 69 (11.7%) | 405 (61.8%) | 527 (47.8%) | 21 (7.6%) | 650 (100.0%) |
| gliner2-fastino-ru | 1235 (32.5%) | 217 (13.3%) | 4 (0.7%) | 82 (12.5%) | 447 (40.5%) | 48 (17.3%) | 640 (98.5%) |
| opf-kz-ru | 1373 (36.2%) | 878 (53.6%) | 7 (1.2%) | 47 (7.2%) | 32 (2.9%) | 240 (86.6%) | 146 (22.5%) |
| ru-pii-ner | 26 (0.7%) | 1186 (72.4%) | 126 (21.4%) | 112 (17.1%) | 830 (75.2%) | 273 (98.6%) | 244 (37.5%) |
| traciora | 488 (12.8%) | 571 (34.9%) | 54 (9.2%) | 257 (39.2%) | 636 (57.7%) | 191 (69.0%) | 633 (97.4%) |
| gliner-pii-edge | 843 (22.2%) | 546 (33.4%) | 65 (11.1%) | 153 (23.4%) | 566 (51.3%) | 63 (22.7%) | 611 (94.0%) |
| fef2-secret-ru | 81 (2.1%) | 361 (22.1%) | 330 (56.1%) | 278 (42.4%) | 1102 (99.9%) | 54 (19.5%) | 645 (99.2%) |
| opf-ru-v2 | 721 (19.0%) | 846 (51.7%) | 36 (6.1%) | 131 (20.0%) | 296 (26.8%) | 197 (71.1%) | 630 (96.9%) |
| gliner2-hivetrace-omni | 934 (24.6%) | 339 (20.7%) | 10 (1.7%) | 294 (44.9%) | 628 (56.9%) | 52 (18.8%) | 644 (99.1%) |
| gliner-nvidia-ru | 936 (24.6%) | 339 (20.7%) | 127 (21.6%) | 153 (23.4%) | 633 (57.4%) | 77 (27.8%) | 644 (99.1%) |
| kalyan-ettin | 1098 (28.9%) | 711 (43.4%) | 137 (23.3%) | 441 (67.3%) | 226 (20.5%) | 173 (62.5%) | 147 (22.6%) |
| stanza-ru | 51 (1.3%) | 344 (21.0%) | 564 (95.9%) | 410 (62.6%) | 926 (84.0%) | 41 (14.8%) | 636 (97.8%) |
| gliner25-fastino | 682 (18.0%) | 524 (32.0%) | 57 (9.7%) | 631 (96.3%) | 306 (27.7%) | 186 (67.1%) | 650 (100.0%) |
| gliner2-hivetrace-omni-ru | 1493 (39.3%) | 353 (21.6%) | 13 (2.2%) | 111 (16.9%) | 620 (56.2%) | 58 (20.9%) | 638 (98.2%) |
| davlan-xlmr | 60 (1.6%) | 225 (13.7%) | 540 (91.8%) | 651 (99.4%) | 1103 (100.0%) | 60 (21.7%) | 650 (100.0%) |
| davlan-mbert | 64 (1.7%) | 328 (20.0%) | 577 (98.1%) | 644 (98.3%) | 1102 (99.9%) | 74 (26.7%) | 650 (100.0%) |
| gliner-multi-v21 | 772 (20.3%) | 477 (29.1%) | 340 (57.8%) | 540 (82.4%) | 670 (60.7%) | 67 (24.2%) | 650 (100.0%) |
| gliner2-hivetrace-uni | 1110 (29.2%) | 526 (32.1%) | 55 (9.4%) | 187 (28.5%) | 956 (86.7%) | 114 (41.2%) | 650 (100.0%) |
| ner-ru-gherman | 152 (4.0%) | 318 (19.4%) | 497 (84.5%) | 654 (99.8%) | 1103 (100.0%) | 248 (89.5%) | 650 (100.0%) |
| spacy-ru-lg | 271 (7.1%) | 439 (26.8%) | 561 (95.4%) | 628 (95.9%) | 1094 (99.2%) | 65 (23.5%) | 648 (99.7%) |
| gliner25-fastino-ru | 1379 (36.3%) | 531 (32.4%) | 107 (18.2%) | 387 (59.1%) | 521 (47.2%) | 219 (79.1%) | 648 (99.7%) |
| gliner2-large | 1793 (47.2%) | 527 (32.2%) | 15 (2.6%) | 256 (39.1%) | 657 (59.6%) | 66 (23.8%) | 608 (93.5%) |
| natasha | 347 (9.1%) | 552 (33.7%) | 588 (100.0%) | 617 (94.2%) | 1101 (99.8%) | 75 (27.1%) | 650 (100.0%) |
| spacy-alrosait | 456 (12.0%) | 319 (19.5%) | 587 (99.8%) | 654 (99.8%) | 1103 (100.0%) | 274 (98.9%) | 650 (100.0%) |
| gliner-urchade-ru | 2531 (66.6%) | 260 (15.9%) | 110 (18.7%) | 142 (21.7%) | 464 (42.1%) | 22 (7.9%) | 635 (97.7%) |
| ner-ru-yqelz | 893 (23.5%) | 529 (32.3%) | 561 (95.4%) | 564 (86.1%) | 982 (89.0%) | 82 (29.6%) | 565 (86.9%) |
| gliner-stream-pii | 1655 (43.6%) | 584 (35.7%) | 242 (41.2%) | 464 (70.8%) | 523 (47.4%) | 136 (49.1%) | 609 (93.7%) |
| gliner-multi-v21-ru | 1586 (41.8%) | 418 (25.5%) | 421 (71.6%) | 419 (64.0%) | 720 (65.3%) | 78 (28.2%) | 650 (100.0%) |
| gliner2-vladlinv | 624 (16.4%) | 1465 (89.5%) | 206 (35.0%) | 346 (52.8%) | 1019 (92.4%) | 275 (99.3%) | 584 (89.8%) |
| gliner2-vladlinv-ru | 1639 (43.2%) | 1480 (90.4%) | 205 (34.9%) | 284 (43.4%) | 1038 (94.1%) | 277 (100.0%) | 591 (90.9%) |
| gravitee-small | 3074 (80.9%) | 1117 (68.2%) | 97 (16.5%) | 422 (64.4%) | 347 (31.5%) | 250 (90.3%) | 271 (41.7%) |
| gliner-pii-base | 3238 (85.3%) | 775 (47.3%) | 108 (18.4%) | 475 (72.5%) | 565 (51.2%) | 77 (27.8%) | 643 (98.9%) |
| gliner2-hivetrace-uni-ru | 2999 (79.0%) | 1198 (73.2%) | 471 (80.1%) | 319 (48.7%) | 1097 (99.5%) | 177 (63.9%) | 650 (100.0%) |
| rules-ru | 3798 (100.0%) | 1630 (99.6%) | 135 (23.0%) | 627 (95.7%) | 329 (29.8%) | 273 (98.6%) | 650 (100.0%) |
| betterleaks | 3798 (100.0%) | 1637 (100.0%) | 588 (100.0%) | 655 (100.0%) | 1103 (100.0%) | 277 (100.0%) | 650 (100.0%) |
| gitleaks | 3798 (100.0%) | 1637 (100.0%) | 588 (100.0%) | 655 (100.0%) | 1103 (100.0%) | 277 (100.0%) | 650 (100.0%) |

## Char recall by gold type

| type | group | nym-base | nym-small | pplx | mmbert32k | nuner-zero | bardsai-eu | gliner2-fastino | apararti | ru-legal-ner+cpu | ru-legal-ner | pii-shield-onnx | opf-ru | openmed-multilingual | gliner-nvidia | openai-base | openmed-nemotron | gliner-urchade | gliner2-fastino-ru | opf-kz-ru | ru-pii-ner | traciora | gliner-pii-edge | fef2-secret-ru | opf-ru-v2 | gliner2-hivetrace-omni | gliner-nvidia-ru | kalyan-ettin | stanza-ru | gliner25-fastino | gliner2-hivetrace-omni-ru | davlan-xlmr | davlan-mbert | gliner-multi-v21 | gliner2-hivetrace-uni | ner-ru-gherman | spacy-ru-lg | gliner25-fastino-ru | gliner2-large | natasha | spacy-alrosait | gliner-urchade-ru | ner-ru-yqelz | gliner-stream-pii | gliner-multi-v21-ru | gliner2-vladlinv | gliner2-vladlinv-ru | gravitee-small | gliner-pii-base | gliner2-hivetrace-uni-ru | rules-ru | betterleaks | gitleaks |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| address | ADDRESS | 0.701 | 0.658 | 0.691 | 0.407 | 0.649 | 0.771 | 0.720 | 0.705 | 0.516 | 0.516 | 0.516 | 0.536 | 0.485 | 0.590 | 0.662 | 0.505 | 0.695 | 0.696 | 0.454 | 0.284 | 0.680 | 0.392 | 0.535 | 0.517 | 0.616 | 0.607 | 0.341 | 0.620 | 0.582 | 0.577 | 0.735 | 0.667 | 0.604 | 0.459 | 0.482 | 0.479 | 0.592 | 0.512 | 0.415 | 0.814 | 0.737 | 0.496 | 0.356 | 0.637 | 0.109 | 0.098 | 0.307 | 0.210 | 0.092 | 0.000 | 0.000 | 0.000 |
| address_apartment | ADDRESS | 0.549 | 0.592 | 1.000 | 0.535 | 0.986 | 0.662 | 0.296 | 0.972 | 0.479 | 0.479 | 0.986 | 0.465 | 0.324 | 0.606 | 0.972 | 0.380 | 0.606 | 0.408 | 0.690 | 0.817 | 0.972 | 0.394 | 0.000 | 0.930 | 0.479 | 0.577 | 0.451 | 0.042 | 0.690 | 0.577 | 0.085 | 0.099 | 0.423 | 0.169 | 0.831 | 0.000 | 0.493 | 0.606 | 0.000 | 0.732 | 0.732 | 0.000 | 0.310 | 0.423 | 0.634 | 0.239 | 0.704 | 0.282 | 0.141 | 0.000 | 0.000 | 0.000 |
| address_building | ADDRESS | 0.236 | 0.327 | 0.655 | 0.400 | 0.636 | 0.582 | 0.400 | 0.945 | 0.509 | 0.509 | 0.873 | 0.418 | 0.000 | 0.891 | 0.800 | 0.400 | 0.455 | 0.400 | 0.709 | 0.218 | 0.982 | 0.455 | 0.055 | 0.709 | 0.218 | 0.218 | 0.200 | 0.655 | 0.473 | 0.218 | 0.055 | 0.055 | 0.236 | 0.018 | 0.418 | 0.109 | 0.382 | 0.273 | 0.109 | 0.564 | 0.291 | 0.018 | 0.073 | 0.236 | 0.218 | 0.218 | 0.091 | 0.073 | 0.000 | 0.000 | 0.000 | 0.000 |
| address_city | ADDRESS | 0.972 | 0.943 | 0.511 | 0.587 | 0.607 | 0.943 | 0.915 | 0.475 | 0.594 | 0.594 | 0.704 | 0.493 | 0.599 | 0.919 | 0.345 | 0.525 | 0.872 | 0.922 | 0.267 | 0.238 | 0.547 | 0.772 | 0.949 | 0.315 | 0.681 | 0.857 | 0.584 | 0.968 | 0.673 | 0.635 | 0.976 | 0.963 | 0.730 | 0.837 | 0.898 | 0.930 | 0.592 | 0.584 | 0.913 | 0.850 | 0.860 | 0.877 | 0.668 | 0.774 | 0.075 | 0.061 | 0.194 | 0.567 | 0.486 | 0.000 | 0.000 | 0.000 |
| address_country | ADDRESS | 0.978 | 0.860 | 0.441 | 0.290 | 0.785 | 1.000 | 1.000 | 0.419 | 0.430 | 0.430 | 0.710 | 0.247 | 0.269 | 0.763 | 0.656 | 0.731 | 0.978 | 1.000 | 0.258 | 0.161 | 0.527 | 0.914 | 0.978 | 0.194 | 0.978 | 0.742 | 0.710 | 1.000 | 0.409 | 0.903 | 1.000 | 1.000 | 0.925 | 0.882 | 0.978 | 0.978 | 0.312 | 0.806 | 1.000 | 0.978 | 0.978 | 0.978 | 0.634 | 0.925 | 0.000 | 0.000 | 0.290 | 0.699 | 0.387 | 0.000 | 0.000 | 0.000 |
| address_district | ADDRESS | 0.694 | 0.668 | 0.318 | 0.252 | 0.366 | 0.821 | 0.753 | 0.501 | 0.505 | 0.505 | 0.645 | 0.393 | 0.382 | 0.601 | 0.301 | 0.330 | 0.762 | 0.740 | 0.250 | 0.328 | 0.539 | 0.494 | 0.823 | 0.386 | 0.650 | 0.583 | 0.323 | 0.870 | 0.267 | 0.688 | 0.800 | 0.833 | 0.240 | 0.796 | 0.355 | 0.814 | 0.306 | 0.529 | 0.729 | 0.723 | 0.750 | 0.749 | 0.515 | 0.528 | 0.101 | 0.060 | 0.132 | 0.210 | 0.362 | 0.000 | 0.000 | 0.000 |
| address_geolocation | ADDRESS | 0.336 | 0.236 | 1.000 | 0.788 | 0.593 | 0.483 | 0.791 | 1.000 | 0.399 | 0.399 | 0.561 | 0.763 | 0.958 | 0.025 | 0.903 | 0.972 | 0.683 | 0.580 | 0.982 | 0.140 | 0.127 | 0.605 | 0.000 | 0.267 | 0.645 | 0.065 | 0.905 | 0.017 | 0.191 | 0.676 | 0.000 | 0.000 | 0.035 | 0.277 | 0.000 | 0.013 | 0.197 | 0.411 | 0.000 | 0.205 | 0.411 | 0.000 | 0.072 | 0.130 | 0.011 | 0.000 | 0.381 | 0.833 | 0.000 | 0.000 | 0.000 | 0.000 |
| address_house | ADDRESS | 0.720 | 0.711 | 0.923 | 0.713 | 0.929 | 0.839 | 0.736 | 0.896 | 0.474 | 0.474 | 0.754 | 0.638 | 0.543 | 0.626 | 0.862 | 0.594 | 0.703 | 0.742 | 0.642 | 0.439 | 0.831 | 0.299 | 0.167 | 0.722 | 0.638 | 0.632 | 0.380 | 0.033 | 0.695 | 0.669 | 0.415 | 0.049 | 0.569 | 0.248 | 0.728 | 0.012 | 0.646 | 0.679 | 0.000 | 0.722 | 0.776 | 0.061 | 0.520 | 0.524 | 0.201 | 0.159 | 0.474 | 0.551 | 0.049 | 0.094 | 0.000 | 0.000 |
| address_postal_code | ADDRESS | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 0.677 | 1.000 | 1.000 | 0.516 | 0.516 | 0.355 | 1.000 | 1.000 | 0.839 | 1.000 | 0.839 | 0.839 | 0.806 | 1.000 | 0.161 | 0.677 | 0.839 | 0.194 | 0.677 | 1.000 | 0.839 | 0.677 | 0.000 | 0.355 | 1.000 | 0.000 | 0.000 | 0.000 | 0.839 | 0.000 | 0.000 | 1.000 | 0.355 | 0.000 | 0.355 | 1.000 | 0.000 | 0.677 | 0.645 | 0.161 | 0.161 | 0.839 | 0.839 | 0.677 | 0.000 | 0.000 | 0.000 |
| address_region | ADDRESS | 0.917 | 0.867 | 0.449 | 0.350 | 0.838 | 0.981 | 0.966 | 0.353 | 0.404 | 0.404 | 0.870 | 0.289 | 0.367 | 0.947 | 0.372 | 0.500 | 1.000 | 0.900 | 0.214 | 0.276 | 0.481 | 0.633 | 0.970 | 0.321 | 0.806 | 0.893 | 0.410 | 0.852 | 0.340 | 0.887 | 0.976 | 0.914 | 0.650 | 0.898 | 0.524 | 0.902 | 0.312 | 0.709 | 0.895 | 0.709 | 0.955 | 0.932 | 0.650 | 0.769 | 0.070 | 0.053 | 0.111 | 0.398 | 0.758 | 0.000 | 0.000 | 0.000 |
| address_street | ADDRESS | 0.881 | 0.822 | 0.759 | 0.438 | 0.832 | 0.898 | 0.925 | 0.741 | 0.603 | 0.603 | 0.524 | 0.697 | 0.586 | 0.696 | 0.678 | 0.614 | 0.861 | 0.873 | 0.449 | 0.296 | 0.756 | 0.411 | 0.632 | 0.532 | 0.761 | 0.745 | 0.303 | 0.801 | 0.799 | 0.750 | 0.964 | 0.900 | 0.858 | 0.542 | 0.554 | 0.580 | 0.837 | 0.654 | 0.468 | 0.869 | 0.911 | 0.553 | 0.409 | 0.870 | 0.128 | 0.140 | 0.298 | 0.135 | 0.022 | 0.000 | 0.000 | 0.000 |
| date | OTHER | 0.926 | 0.892 | 0.753 | 0.685 | 0.010 | 0.025 | 0.008 | 0.762 | 0.737 | 0.737 | 0.589 | 0.307 | 0.643 | 0.005 | 0.658 | 0.557 | 0.000 | 0.014 | 0.724 | 0.648 | 0.015 | 0.040 | 0.005 | 0.018 | 0.006 | 0.008 | 0.639 | 0.020 | 0.000 | 0.014 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.003 | 0.003 | 0.053 | 0.000 | 0.000 | 0.021 | 0.120 | 0.025 | 0.000 | 0.100 | 0.092 | 0.500 | 0.007 | 0.000 | 0.000 | 0.000 | 0.000 |
| document_number | ID | 0.720 | 0.714 | 0.879 | 0.767 | 0.688 | 0.774 | 0.720 | 0.919 | 0.736 | 0.736 | 0.628 | 0.680 | 0.756 | 0.626 | 0.855 | 0.314 | 0.230 | 0.908 | 0.913 | 0.809 | 0.576 | 0.638 | 0.498 | 0.707 | 0.443 | 0.784 | 0.176 | 0.173 | 0.049 | 0.801 | 0.002 | 0.003 | 0.145 | 0.659 | 0.001 | 0.059 | 0.445 | 0.508 | 0.032 | 0.003 | 0.895 | 0.033 | 0.208 | 0.368 | 0.433 | 0.569 | 0.376 | 0.135 | 0.425 | 0.048 | 0.000 | 0.000 |
| email | CONTACT | 0.935 | 0.948 | 0.999 | 0.848 | 0.999 | 0.719 | 0.997 | 0.975 | 0.906 | 0.906 | 0.990 | 0.917 | 0.970 | 0.935 | 0.948 | 0.987 | 0.962 | 0.999 | 0.966 | 0.681 | 0.982 | 0.845 | 0.968 | 0.859 | 0.999 | 0.923 | 0.989 | 0.105 | 0.915 | 0.999 | 0.117 | 0.031 | 0.271 | 0.979 | 0.169 | 0.121 | 0.679 | 0.995 | 0.000 | 0.005 | 0.880 | 0.026 | 0.644 | 0.216 | 0.636 | 0.626 | 0.740 | 0.979 | 0.522 | 0.997 | 0.000 | 0.000 |
| first_name | PERSON | 0.993 | 0.948 | 0.992 | 0.832 | 0.945 | 0.987 | 0.966 | 0.666 | 0.825 | 0.825 | 0.700 | 0.852 | 0.686 | 0.953 | 0.665 | 0.755 | 0.870 | 0.623 | 0.575 | 0.998 | 0.804 | 0.763 | 0.970 | 0.670 | 0.754 | 0.717 | 0.758 | 0.980 | 0.885 | 0.625 | 0.989 | 0.988 | 0.850 | 0.774 | 0.986 | 0.942 | 0.650 | 0.578 | 0.921 | 0.881 | 0.216 | 0.748 | 0.613 | 0.620 | 0.782 | 0.529 | 0.189 | 0.251 | 0.196 | 0.000 | 0.000 | 0.000 |
| ip | NET | 0.663 | 0.662 | 0.957 | 0.591 | 0.956 | 0.472 | 0.400 | 0.947 | 0.255 | 0.255 | 0.769 | 0.281 | 0.906 | 0.432 | 0.805 | 0.712 | 0.415 | 0.402 | 0.930 | 0.254 | 0.283 | 0.415 | 0.000 | 0.549 | 0.294 | 0.345 | 0.669 | 0.058 | 0.730 | 0.298 | 0.000 | 0.000 | 0.269 | 0.095 | 0.000 | 0.007 | 0.544 | 0.275 | 0.001 | 0.000 | 0.463 | 0.063 | 0.313 | 0.236 | 0.079 | 0.062 | 0.501 | 0.361 | 0.003 | 0.620 | 0.000 | 0.000 |
| last_name | PERSON | 0.998 | 0.986 | 0.988 | 0.783 | 0.896 | 0.990 | 0.918 | 0.798 | 0.966 | 0.966 | 0.812 | 0.849 | 0.591 | 0.891 | 0.752 | 0.752 | 0.910 | 0.684 | 0.582 | 0.997 | 0.882 | 0.765 | 0.990 | 0.798 | 0.680 | 0.762 | 0.567 | 0.991 | 0.741 | 0.567 | 0.990 | 0.979 | 0.773 | 0.633 | 0.997 | 0.939 | 0.605 | 0.507 | 0.909 | 0.903 | 0.323 | 0.734 | 0.523 | 0.584 | 0.848 | 0.581 | 0.145 | 0.075 | 0.208 | 0.000 | 0.000 | 0.000 |
| middle_name | PERSON | 1.000 | 0.998 | 1.000 | 0.775 | 0.923 | 1.000 | 0.951 | 0.942 | 1.000 | 1.000 | 0.921 | 0.992 | 0.708 | 0.695 | 0.915 | 0.850 | 0.781 | 0.579 | 0.846 | 1.000 | 0.984 | 0.739 | 0.992 | 0.955 | 0.670 | 0.433 | 0.681 | 1.000 | 0.822 | 0.480 | 0.990 | 0.994 | 0.798 | 0.696 | 0.996 | 0.944 | 0.744 | 0.463 | 0.918 | 0.890 | 0.199 | 0.848 | 0.541 | 0.587 | 0.829 | 0.552 | 0.264 | 0.016 | 0.179 | 0.000 | 0.000 | 0.000 |
| military_individual_number | ID | 0.579 | 0.551 | 0.942 | 0.583 | 0.872 | 0.656 | 0.891 | 0.975 | 0.510 | 0.510 | 0.483 | 0.424 | 0.784 | 0.874 | 0.897 | 0.207 | 0.646 | 0.864 | 0.947 | 0.821 | 0.319 | 0.894 | 0.860 | 0.513 | 0.909 | 0.572 | 0.335 | 0.226 | 0.024 | 0.700 | 0.019 | 0.012 | 0.250 | 0.766 | 0.000 | 0.041 | 0.093 | 0.704 | 0.146 | 0.000 | 0.612 | 0.146 | 0.364 | 0.321 | 0.732 | 0.581 | 0.130 | 0.572 | 0.809 | 0.000 | 0.000 | 0.000 |
| mobile_phone | CONTACT | 0.982 | 0.985 | 0.998 | 0.933 | 0.972 | 0.940 | 0.987 | 0.997 | 0.956 | 0.956 | 0.959 | 0.954 | 0.947 | 0.742 | 0.991 | 0.780 | 0.848 | 0.989 | 0.993 | 0.866 | 0.858 | 0.909 | 0.091 | 0.967 | 0.976 | 0.688 | 0.519 | 0.002 | 0.902 | 0.969 | 0.000 | 0.000 | 0.421 | 0.865 | 0.000 | 0.002 | 0.889 | 0.961 | 0.000 | 0.000 | 0.782 | 0.016 | 0.527 | 0.252 | 0.691 | 0.699 | 0.814 | 0.756 | 0.019 | 0.686 | 0.000 | 0.000 |
| name | PERSON | 0.949 | 0.953 | 0.994 | 0.748 | 0.915 | 0.995 | 0.948 | 0.891 | 0.976 | 0.976 | 0.813 | 0.886 | 0.646 | 0.830 | 0.872 | 0.760 | 0.804 | 0.629 | 0.715 | 0.994 | 0.932 | 0.719 | 0.986 | 0.908 | 0.717 | 0.641 | 0.662 | 0.996 | 0.767 | 0.520 | 0.984 | 0.988 | 0.734 | 0.661 | 0.899 | 0.926 | 0.644 | 0.427 | 0.904 | 0.885 | 0.348 | 0.775 | 0.470 | 0.528 | 0.854 | 0.587 | 0.170 | 0.047 | 0.183 | 0.000 | 0.000 | 0.000 |
| name_initials | PERSON | 0.914 | 0.966 | 0.977 | 0.348 | 0.810 | 0.981 | 0.797 | 0.787 | 0.825 | 0.825 | 0.365 | 0.498 | 0.215 | 0.703 | 0.827 | 0.278 | 0.441 | 0.378 | 0.061 | 0.973 | 0.477 | 0.483 | 0.973 | 0.759 | 0.593 | 0.380 | 0.165 | 0.998 | 0.430 | 0.359 | 0.968 | 0.977 | 0.175 | 0.538 | 0.114 | 0.732 | 0.222 | 0.329 | 0.844 | 0.722 | 0.635 | 0.597 | 0.224 | 0.110 | 0.819 | 0.542 | 0.065 | 0.000 | 0.424 | 0.000 | 0.000 | 0.000 |
| nickname | PERSON | 0.798 | 0.532 | 0.972 | 0.355 | 0.968 | 0.681 | 0.957 | 0.199 | 0.323 | 0.323 | 0.113 | 0.518 | 0.411 | 0.972 | 0.252 | 0.330 | 0.972 | 0.872 | 0.181 | 0.638 | 0.372 | 0.780 | 0.273 | 0.199 | 0.947 | 0.897 | 0.248 | 0.504 | 0.801 | 0.972 | 0.355 | 0.305 | 0.613 | 0.940 | 0.142 | 0.443 | 0.709 | 0.908 | 0.092 | 0.014 | 0.897 | 0.202 | 0.220 | 0.667 | 0.631 | 0.589 | 0.096 | 0.613 | 0.879 | 0.000 | 0.000 | 0.000 |
| organization | ORG | 0.710 | 0.675 | 0.066 | 0.352 | 0.844 | 0.813 | 0.834 | 0.315 | 0.601 | 0.601 | 0.148 | 0.238 | 0.215 | 0.823 | 0.215 | 0.282 | 0.938 | 0.831 | 0.101 | 0.007 | 0.280 | 0.717 | 0.798 | 0.246 | 0.806 | 0.720 | 0.300 | 0.803 | 0.329 | 0.805 | 0.803 | 0.752 | 0.783 | 0.535 | 0.053 | 0.775 | 0.212 | 0.753 | 0.724 | 0.012 | 0.927 | 0.717 | 0.478 | 0.751 | 0.003 | 0.000 | 0.101 | 0.705 | 0.294 | 0.018 | 0.000 | 0.000 |
| snils | ID | 0.956 | 0.955 | 0.944 | 0.856 | 0.639 | 0.728 | 0.928 | 0.966 | 0.813 | 0.813 | 0.888 | 0.916 | 0.886 | 0.679 | 0.855 | 0.478 | 0.249 | 0.963 | 0.957 | 0.947 | 0.735 | 0.539 | 0.581 | 0.909 | 0.361 | 0.811 | 0.145 | 0.039 | 0.066 | 0.942 | 0.000 | 0.010 | 0.225 | 0.899 | 0.000 | 0.007 | 0.859 | 0.904 | 0.000 | 0.000 | 0.970 | 0.036 | 0.214 | 0.341 | 0.635 | 0.893 | 0.617 | 0.103 | 0.346 | 0.019 | 0.000 | 0.000 |
| tin | ID | 0.938 | 0.938 | 0.945 | 0.913 | 0.629 | 0.919 | 0.782 | 0.951 | 0.877 | 0.877 | 0.850 | 0.909 | 0.936 | 0.716 | 0.896 | 0.316 | 0.614 | 0.938 | 0.960 | 0.837 | 0.828 | 0.805 | 0.341 | 0.890 | 0.443 | 0.915 | 0.358 | 0.536 | 0.025 | 0.938 | 0.000 | 0.011 | 0.294 | 0.523 | 0.000 | 0.000 | 0.580 | 0.184 | 0.000 | 0.000 | 0.915 | 0.044 | 0.165 | 0.574 | 0.475 | 0.297 | 0.165 | 0.051 | 0.405 | 0.152 | 0.000 | 0.000 |
| vehicle_number | ID | 0.962 | 0.950 | 0.975 | 0.811 | 0.866 | 0.918 | 0.965 | 0.926 | 0.635 | 0.635 | 0.668 | 0.272 | 0.753 | 0.852 | 0.770 | 0.231 | 0.879 | 0.571 | 0.844 | 0.660 | 0.308 | 0.879 | 0.454 | 0.846 | 0.925 | 0.836 | 0.332 | 0.969 | 0.000 | 0.934 | 0.003 | 0.000 | 0.053 | 0.805 | 0.000 | 0.000 | 0.013 | 0.909 | 0.052 | 0.000 | 0.112 | 0.509 | 0.170 | 0.143 | 0.228 | 0.362 | 0.046 | 0.807 | 0.830 | 0.000 | 0.000 | 0.000 |

## char F1 by domain

| domain | n | nym-base | nym-small | pplx | mmbert32k | nuner-zero | bardsai-eu | gliner2-fastino | apararti | ru-legal-ner+cpu | ru-legal-ner | pii-shield-onnx | opf-ru | openmed-multilingual | gliner-nvidia | openai-base | openmed-nemotron | gliner-urchade | gliner2-fastino-ru | opf-kz-ru | ru-pii-ner | traciora | gliner-pii-edge | fef2-secret-ru | opf-ru-v2 | gliner2-hivetrace-omni | gliner-nvidia-ru | kalyan-ettin | stanza-ru | gliner25-fastino | gliner2-hivetrace-omni-ru | davlan-xlmr | davlan-mbert | gliner-multi-v21 | gliner2-hivetrace-uni | ner-ru-gherman | spacy-ru-lg | gliner25-fastino-ru | gliner2-large | natasha | spacy-alrosait | gliner-urchade-ru | ner-ru-yqelz | gliner-stream-pii | gliner-multi-v21-ru | gliner2-vladlinv | gliner2-vladlinv-ru | gravitee-small | gliner-pii-base | gliner2-hivetrace-uni-ru | rules-ru | betterleaks | gitleaks |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| controlled | 1500 | 0.856 | 0.846 | 0.862 | 0.758 | 0.803 | 0.801 | 0.781 | 0.817 | 0.743 | 0.743 | 0.789 | 0.712 | 0.775 | 0.718 | 0.775 | 0.691 | 0.696 | 0.719 | 0.760 | 0.707 | 0.685 | 0.666 | 0.680 | 0.715 | 0.695 | 0.682 | 0.698 | 0.624 | 0.723 | 0.664 | 0.636 | 0.622 | 0.600 | 0.632 | 0.549 | 0.589 | 0.660 | 0.576 | 0.566 | 0.581 | 0.614 | 0.531 | 0.539 | 0.545 | 0.558 | 0.473 | 0.460 | 0.447 | 0.286 | 0.364 | 0.000 | 0.000 |

## Missed at thresholds 0.5 / 0.3 / 0.2 / 0.1

Only models that return a score. A lower threshold keeps more spans: fewer misses, more masking.

| model | 0.5 | 0.3 | 0.2 | 0.1 |
|---|---|---|---|---|
| nym-base | 776 (8.9%) | 710 (8.2%) | 709 (8.1%) | 709 (8.1%) |
| nym-small | 904 (10.4%) | 826 (9.5%) | 823 (9.5%) | 823 (9.5%) |
| mmbert32k | 1413 (16.2%) | 536 (6.2%) | 468 (5.4%) | 462 (5.3%) |
| nuner-zero | 1529 (17.6%) | 1004 (11.5%) | 830 (9.5%) | 695 (8.0%) |
| bardsai-eu | 1591 (18.3%) | 1395 (16.0%) | 1387 (15.9%) | 1386 (15.9%) |
| gliner2-fastino | 1638 (18.8%) | 1361 (15.6%) | 1257 (14.4%) | 1137 (13.1%) |
| apararti | 1642 (18.9%) | 1601 (18.4%) | 1598 (18.4%) | 1598 (18.4%) |
| ru-legal-ner+cpu | 1724 (19.8%) | 885 (10.2%) | 775 (8.9%) | 759 (8.7%) |
| ru-legal-ner | 1724 (19.8%) | 885 (10.2%) | 775 (8.9%) | 759 (8.7%) |
| opf-ru | 2044 (23.5%) | 1935 (22.2%) | 1929 (22.2%) | 1929 (22.2%) |
| openmed-multilingual | 2071 (23.8%) | 1657 (19.0%) | 1633 (18.8%) | 1632 (18.7%) |
| gliner-nvidia | 2172 (24.9%) | 1678 (19.3%) | 1452 (16.7%) | 1199 (13.8%) |
| openai-base | 2230 (25.6%) | 2218 (25.5%) | 2218 (25.5%) | 2218 (25.5%) |
| openmed-nemotron | 2364 (27.1%) | 1940 (22.3%) | 1915 (22.0%) | 1915 (22.0%) |
| gliner-urchade | 2389 (27.4%) | 1787 (20.5%) | 1537 (17.7%) | 1290 (14.8%) |
| gliner2-fastino-ru | 2673 (30.7%) | 1726 (19.8%) | 1454 (16.7%) | 1207 (13.9%) |
| opf-kz-ru | 2723 (31.3%) | 2703 (31.0%) | 2703 (31.0%) | 2703 (31.0%) |
| traciora | 2830 (32.5%) | 2808 (32.2%) | 2806 (32.2%) | 2806 (32.2%) |
| gliner-pii-edge | 2847 (32.7%) | 554 (6.4%) | 374 (4.3%) | 230 (2.6%) |
| fef2-secret-ru | 2851 (32.7%) | 2813 (32.3%) | 2809 (32.3%) | 2809 (32.3%) |
| opf-ru-v2 | 2857 (32.8%) | 2829 (32.5%) | 2828 (32.5%) | 2828 (32.5%) |
| gliner2-hivetrace-omni | 2901 (33.3%) | 2249 (25.8%) | 1899 (21.8%) | 1569 (18.0%) |
| gliner-nvidia-ru | 2909 (33.4%) | 2124 (24.4%) | 1764 (20.3%) | 1370 (15.7%) |
| kalyan-ettin | 2933 (33.7%) | 2355 (27.0%) | 2294 (26.3%) | 2286 (26.3%) |
| gliner25-fastino | 3036 (34.9%) | 2584 (29.7%) | 2383 (27.4%) | 2182 (25.1%) |
| gliner2-hivetrace-omni-ru | 3286 (37.7%) | 2242 (25.7%) | 1826 (21.0%) | 1416 (16.3%) |
| davlan-xlmr | 3289 (37.8%) | 3287 (37.7%) | 3287 (37.7%) | 3287 (37.7%) |
| davlan-mbert | 3439 (39.5%) | 3437 (39.5%) | 3437 (39.5%) | 3437 (39.5%) |
| gliner-multi-v21 | 3516 (40.4%) | 2101 (24.1%) | 1512 (17.4%) | 1065 (12.2%) |
| gliner2-hivetrace-uni | 3598 (41.3%) | 2299 (26.4%) | 1783 (20.5%) | 1217 (14.0%) |
| ner-ru-gherman | 3622 (41.6%) | 3545 (40.7%) | 3536 (40.6%) | 3536 (40.6%) |
| gliner25-fastino-ru | 3792 (43.5%) | 3209 (36.9%) | 2940 (33.8%) | 2706 (31.1%) |
| gliner2-large | 3922 (45.0%) | 2451 (28.1%) | 1873 (21.5%) | 1384 (15.9%) |
| gliner-urchade-ru | 4164 (47.8%) | 3220 (37.0%) | 2740 (31.5%) | 2039 (23.4%) |
| ner-ru-yqelz | 4176 (48.0%) | 4070 (46.7%) | 4070 (46.7%) | 4070 (46.7%) |
| gliner-stream-pii | 4213 (48.4%) | 2994 (34.4%) | 2373 (27.3%) | 1557 (17.9%) |
| gliner-multi-v21-ru | 4292 (49.3%) | 2323 (26.7%) | 1592 (18.3%) | 1009 (11.6%) |
| gliner2-vladlinv | 4519 (51.9%) | 4408 (50.6%) | 4332 (49.7%) | 4224 (48.5%) |
| gliner2-vladlinv-ru | 5514 (63.3%) | 5378 (61.8%) | 5312 (61.0%) | 5227 (60.0%) |
| gravitee-small | 5578 (64.1%) | 5424 (62.3%) | 5422 (62.3%) | 5422 (62.3%) |
| gliner-pii-base | 5881 (67.5%) | 1469 (16.9%) | 777 (8.9%) | 420 (4.8%) |
| gliner2-hivetrace-uni-ru | 6911 (79.4%) | 4837 (55.5%) | 3605 (41.4%) | 1962 (22.5%) |

Notes: the unit of counting is one gold span after boundary normalization - the same unit in `missed`, in the group table, in the threshold table and in the ensembles; two annotations that share boundaries but not type are two spans, so an empty answer misses 100%. Gold and predicted boundaries are normalized the same way (whitespace, word boundaries, merge of adjacent same-label pieces). `dropped spans` - predicted intervals outside the text: rejected, not clipped to fit. A run made before `run.py` started clipping span ends to the piece it fed the model can carry them; `meta.clipped` counts what the clipping fixes in newer runs. Zero-shot models get the taxonomy of the set as labels (the `labels` field of `meta.json`). `train` - the model was trained on the source of this set (its slices and corrupted copies included): its row is not comparable with the others and stays out of the pooled numbers; an empty `contaminated` list in `benchmark/models.toml` means no evidence of overlap was found, not proof of none. `hidden` - gold spans every character of which is covered; a touched span counts as detected, not as hidden. missed % and char F1 carry 95% bootstrap intervals over rows (1000 resamples, the same rows for every model of the set). The group of a gold type comes from `BENCH/<set>/meta.json`; a model label with no group falls into OTHER. ms/row is a median and includes chunking of long texts; hardware and batch are in the meta line of the prediction file.