Document the 4th (vision-based) extractor family

README.md

@@ -41,19 +41,27 @@ Code, loaders, scoring harness, and both papers' reproducibility artifacts:
 
 ## What's in this repository
 
-Eight genuine per-field capture files, **29,755 records total**, from three extractor
+Nine genuine per-field capture files, **36,265 records total**, from four extractor
 families over four corpora:
 
 | file | extractor | fields | docs | source corpus |
 |---|---|---|---|---|
-| `cord_claude-sonnet-5.json` | claude-sonnet-5 | 13,859 | 800 | CORD (receipts) |
-| `funsd_claude-sonnet-5.json` | claude-sonnet-5 | 1,999 | 175 | FUNSD (forms) |
-| `xfund_de_claude-sonnet-5.json` | claude-sonnet-5 | 523 | 42 | XFUND German |
-| `xfund_es_claude-sonnet-5.json` | claude-sonnet-5 | 632 | 43 | XFUND Spanish |
-| `xfund_fr_claude-sonnet-5.json` | claude-sonnet-5 | 772 | 44 | XFUND French |
-| `xfund_zh_claude-sonnet-5.json` | claude-sonnet-5 | 461 | 27 | XFUND Chinese |
-| `cord_claude-haiku-4-5.json` | claude-haiku-4-5 | 5,341 | 400 | CORD (frozen-config confirmation) |
-| `cord_qwen2.5-14b.json` | Qwen2.5-14B (vLLM, open-weights) | 6,168 | 398 | CORD |
+| `cord_claude-sonnet-5.json` | claude-sonnet-5 (text-layer) | 13,859 | 800 | CORD (receipts) |
+| `funsd_claude-sonnet-5.json` | claude-sonnet-5 (text-layer) | 1,999 | 175 | FUNSD (forms) |
+| `xfund_de_claude-sonnet-5.json` | claude-sonnet-5 (text-layer) | 523 | 42 | XFUND German |
+| `xfund_es_claude-sonnet-5.json` | claude-sonnet-5 (text-layer) | 632 | 43 | XFUND Spanish |
+| `xfund_fr_claude-sonnet-5.json` | claude-sonnet-5 (text-layer) | 772 | 44 | XFUND French |
+| `xfund_zh_claude-sonnet-5.json` | claude-sonnet-5 (text-layer) | 461 | 27 | XFUND Chinese |
+| `cord_claude-haiku-4-5.json` | claude-haiku-4-5 (text-layer) | 5,341 | 400 | CORD (frozen-config confirmation) |
+| `cord_qwen2.5-14b.json` | Qwen2.5-14B (vLLM, open-weights, text-layer) | 6,168 | 398 | CORD |
+| `cord_gpt-4o.json` | **gpt-4o (genuinely vision-based)** | 6,510 | 399 | CORD |
+
+**Only `cord_gpt-4o.json` is genuinely vision-based** — the model was sent the
+rendered page image (base64-encoded) alongside the OCR text layer as assistive
+context. Every other file is **text-layer prompted**: the model reads only the
+document's OCR text, never pixels, despite "VLM" being the natural shorthand for a
+closed frontier API model — we correct that framing here rather than let it stand
+for three of the four extractor families.
 
 Each record is one field prediction: