Datasets:
Release: data, probe and finetune sets, recipes, analysis outputs (1/1)
Browse files
data/probe_sets/corpus_starts_v4.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"what": "how the corpus's bad payoffs start, per performed act (scripts/paper_tables.py corpus_starts)",
|
| 3 |
+
"sources": [
|
| 4 |
+
"data/raw/v4-one/bundle5.jsonl",
|
| 5 |
+
"data/raw/v4-one/bundle5b.jsonl",
|
| 6 |
+
"data/raw/v4-both/bundle5.jsonl",
|
| 7 |
+
"data/raw/v4-both/bundle5b.jsonl"
|
| 8 |
+
],
|
| 9 |
+
"starts": {
|
| 10 |
+
"tease": {
|
| 11 |
+
"quote": 559075,
|
| 12 |
+
"name": 701,
|
| 13 |
+
"other": 63
|
| 14 |
+
},
|
| 15 |
+
"steal": {
|
| 16 |
+
"while": 455693,
|
| 17 |
+
"name": 101739,
|
| 18 |
+
"quote": 1202,
|
| 19 |
+
"other": 427
|
| 20 |
+
},
|
| 21 |
+
"cheat": {
|
| 22 |
+
"while": 225872,
|
| 23 |
+
"name": 332456,
|
| 24 |
+
"other": 1766,
|
| 25 |
+
"quote": 554
|
| 26 |
+
},
|
| 27 |
+
"lie": {
|
| 28 |
+
"quote": 560333,
|
| 29 |
+
"other": 104,
|
| 30 |
+
"name": 14,
|
| 31 |
+
"while": 1
|
| 32 |
+
}
|
| 33 |
+
},
|
| 34 |
+
"n": {
|
| 35 |
+
"tease": 559839,
|
| 36 |
+
"steal": 559061,
|
| 37 |
+
"cheat": 560648,
|
| 38 |
+
"lie": 560452
|
| 39 |
+
}
|
| 40 |
+
}
|