ab_non_ppsn_v5_1.json

{
  "base_model": "OpenMed/OpenMed-PII-SuperClinical-Large-434M-v1",
  "candidate_model": "models/openmed-ppsns-lora-v5_1",
  "dataset": {
    "synthetic_source": "nvidia/Nemotron-PII:test",
    "real_source": "wikitext/wikitext-2-raw-v1:validation",
    "synthetic_size": 400,
    "real_size": 400,
    "match_iou_threshold": 0.5
  },
  "synthetic_non_ppsn_vs_gold": {
    "base": {
      "precision": 0.8194517659462309,
      "recall": 0.9275059665871122,
      "f1": 0.8701371396585503,
      "tp": 3109,
      "fp": 685,
      "fn": 243
    },
    "candidate": {
      "precision": 0.8203166226912929,
      "recall": 0.9275059665871122,
      "f1": 0.8706244749369925,
      "tp": 3109,
      "fp": 681,
      "fn": 243
    }
  },
  "real_non_ppsn_behavior": {
    "base": {
      "docs": 400,
      "total_non_ppsn_entities": 2643,
      "entities_per_1k_chars": 11.412409862256574,
      "docs_with_any_entity": 339,
      "docs_with_any_entity_rate": 0.8475,
      "top_labels": [
        [
          "last_name",
          797
        ],
        [
          "city",
          516
        ],
        [
          "first_name",
          455
        ],
        [
          "state",
          266
        ],
        [
          "date",
          204
        ],
        [
          "street_address",
          88
        ],
        [
          "occupation",
          85
        ],
        [
          "country",
          74
        ],
        [
          "company_name",
          41
        ],
        [
          "county",
          40
        ],
        [
          "race_ethnicity",
          31
        ],
        [
          "language",
          9
        ],
        [
          "date_of_birth",
          9
        ],
        [
          "education_level",
          7
        ],
        [
          "password",
          6
        ]
      ]
    },
    "candidate": {
      "docs": 400,
      "total_non_ppsn_entities": 2643,
      "entities_per_1k_chars": 11.412409862256574,
      "docs_with_any_entity": 339,
      "docs_with_any_entity_rate": 0.8475,
      "top_labels": [
        [
          "last_name",
          797
        ],
        [
          "city",
          516
        ],
        [
          "first_name",
          455
        ],
        [
          "state",
          266
        ],
        [
          "date",
          204
        ],
        [
          "street_address",
          88
        ],
        [
          "occupation",
          85
        ],
        [
          "country",
          74
        ],
        [
          "company_name",
          41
        ],
        [
          "county",
          40
        ],
        [
          "race_ethnicity",
          31
        ],
        [
          "language",
          9
        ],
        [
          "date_of_birth",
          9
        ],
        [
          "education_level",
          7
        ],
        [
          "password",
          6
        ]
      ]
    },
    "candidate_vs_base_agreement": {
      "precision": 1.0,
      "recall": 1.0,
      "f1": 1.0,
      "tp": 2643,
      "fp": 0,
      "fn": 0
    }
  }
}