raw-extraction build (model predictions, no judge)

README.md

@@ -0,0 +1,40 @@
+---
+task_categories:
+  - token-classification
+tags:
+  - gliner
+  - gliner2
+  - information-extraction
+  - probe-review-sample
+license: cc-by-4.0
+configs:
+  - config_name: gliner
+    data_files:
+      - split: sample
+        path: gliner_sample.jsonl
+  - config_name: bio
+    data_files:
+      - split: sample
+        path: bio_sample.jsonl
+  - config_name: gliner2
+    data_files:
+      - split: sample
+        path: gliner2_sample.jsonl
+---
+
+# Probe-review sample of `rafmacalaba/datause-extracted-sample`
+
+Stratified review slice: whole rows sampled per origin until each specificity reaches ~150 spans (seed 0; small origins contribute all they have), plus entity-less calibration rows round-robined across `neg_class` shapes. Rows keep their original `split` values. Score with training/score_extract_probe.py `--splits sample`.
+
+| origin | rows | negatives | named / descriptive / vague spans |
+|---|---|---|---|
+| `fcv_pads_east_africa` | 389 | 60 | 171 / 187 / 150 |
+| `general_prwp` | 317 | 60 | 192 / 187 / 150 |
+| `jad_paddy_docs` | 140 | 7 | 187 / 63 / 46 |
+| `jdc_operational` | 310 | 60 | 175 / 157 / 76 |
+| `refugee_pads` | 376 | 60 | 169 / 182 / 150 |
+| `reliefweb` | 339 | 60 | 166 / 170 / 150 |
+
+## Probe scores
+
+Every span carries `probe_score` (sigmoid head output ~= P(genuine data-use mention)) from `rafmacalaba/gliner_datause_probe` (frozen `rafmacalaba/gliner_datause` encoder, context_radius=64, threshold=None). Extractor proposes, head disposes — see `probe.json`.

labels.json

@@ -0,0 +1,25 @@
+{
+  "gliner_labels": [
+    "NAMED_DATA",
+    "DESCRIPTIVE_DATA",
+    "VAGUE_DATA"
+  ],
+  "bio_tags": [
+    "O",
+    "B-NAMED_DATA",
+    "I-NAMED_DATA",
+    "B-DESCRIPTIVE_DATA",
+    "I-DESCRIPTIVE_DATA",
+    "B-VAGUE_DATA",
+    "I-VAGUE_DATA"
+  ],
+  "label_map": {
+    "O": 0,
+    "B-NAMED_DATA": 1,
+    "I-NAMED_DATA": 2,
+    "B-DESCRIPTIVE_DATA": 3,
+    "I-DESCRIPTIVE_DATA": 4,
+    "B-VAGUE_DATA": 5,
+    "I-VAGUE_DATA": 6
+  }
+}

probe.json

@@ -0,0 +1,9 @@
+{
+  "probe_model": "rafmacalaba/gliner_datause_probe",
+  "base_encoder": "rafmacalaba/gliner_datause",
+  "threshold": null,
+  "date": "2026-09-06",
+  "input_dim": 2048,
+  "context_radius": 64,
+  "hidden_dim": 256
+}

split_stats.json

@@ -0,0 +1,62 @@
+{
+  "sample": {
+    "origins": {
+      "fcv_pads_east_africa": {
+        "spans": {
+          "VAGUE_DATA": 150,
+          "NAMED_DATA": 171,
+          "DESCRIPTIVE_DATA": 187
+        },
+        "rows": 389,
+        "negatives": 60
+      },
+      "general_prwp": {
+        "spans": {
+          "DESCRIPTIVE_DATA": 187,
+          "VAGUE_DATA": 150,
+          "NAMED_DATA": 192
+        },
+        "rows": 317,
+        "negatives": 60
+      },
+      "jad_paddy_docs": {
+        "spans": {
+          "NAMED_DATA": 187,
+          "DESCRIPTIVE_DATA": 63,
+          "VAGUE_DATA": 46
+        },
+        "rows": 140,
+        "negatives": 7
+      },
+      "jdc_operational": {
+        "spans": {
+          "NAMED_DATA": 175,
+          "DESCRIPTIVE_DATA": 157,
+          "VAGUE_DATA": 76
+        },
+        "rows": 310,
+        "negatives": 60
+      },
+      "refugee_pads": {
+        "spans": {
+          "NAMED_DATA": 169,
+          "DESCRIPTIVE_DATA": 182,
+          "VAGUE_DATA": 150
+        },
+        "rows": 376,
+        "negatives": 60
+      },
+      "reliefweb": {
+        "spans": {
+          "NAMED_DATA": 166,
+          "DESCRIPTIVE_DATA": 170,
+          "VAGUE_DATA": 150
+        },
+        "rows": 339,
+        "negatives": 60
+      }
+    },
+    "per_cell": 150,
+    "seed": 0
+  }
+}