rafmacalaba commited on
Commit
94bcd10
·
verified ·
1 Parent(s): 459ad82

FCV data-use paper corpus (6 configs)

Browse files
README.md CHANGED
@@ -80,8 +80,7 @@ All splits are published verbatim exactly as used in the paper.
80
  ## Schemas
81
 
82
  - `extraction` / `synthetic` / `sources`: `{"input", "output": {"entities":
83
- {"named_data", "descriptive_data", "vague_data"}, "entity_descriptions"}}`;
84
- `extraction/test` additionally carries the model's `predictions`.
85
  - `extraction` rows carry a `meta` field tagging provenance: `FCV`
86
  (forced-displacement / FCV corpus) or `PRWP` (Policy Research Working Papers
87
  -- manually annotated + reviewed/validated). `train` = FCV (2,473) + PRWP
 
80
  ## Schemas
81
 
82
  - `extraction` / `synthetic` / `sources`: `{"input", "output": {"entities":
83
+ {"named_data", "descriptive_data", "vague_data"}, "entity_descriptions"}}`.
 
84
  - `extraction` rows carry a `meta` field tagging provenance: `FCV`
85
  (forced-displacement / FCV corpus) or `PRWP` (Policy Research Working Papers
86
  -- manually annotated + reviewed/validated). `train` = FCV (2,473) + PRWP
extraction/test.jsonl CHANGED
The diff for this file is too large to render. See raw diff
 
extraction/train.jsonl CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:fc15491ebed5c6d0f8981077710322282adcab2f189cf63852e07ae19fca328e
3
- size 10937882
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:246b973a130f61e95d909302eb62ca8af82b9b7cc58203202dfc95e90f1979d6
3
+ size 10466229
extraction/validation.jsonl CHANGED
The diff for this file is too large to render. See raw diff