albertvillanova HF staff commited on
Commit
c43d022
1 Parent(s): 094ce12

Add principle_A_reconstruction data files

Browse files
README.md CHANGED
@@ -1388,10 +1388,10 @@ dataset_info:
1388
  dtype: int32
1389
  splits:
1390
  - name: train
1391
- num_bytes: 153499
1392
  num_examples: 1000
1393
- download_size: 345494
1394
- dataset_size: 153499
1395
  - config_name: regular_plural_subject_verb_agreement_1
1396
  features:
1397
  - name: sentence_good
@@ -2089,6 +2089,10 @@ configs:
2089
  data_files:
2090
  - split: train
2091
  path: principle_A_domain_3/train-*
 
 
 
 
2092
  ---
2093
 
2094
  # Dataset Card for "blimp"
 
1388
  dtype: int32
1389
  splits:
1390
  - name: train
1391
+ num_bytes: 152104
1392
  num_examples: 1000
1393
+ download_size: 44480
1394
+ dataset_size: 152104
1395
  - config_name: regular_plural_subject_verb_agreement_1
1396
  features:
1397
  - name: sentence_good
 
2089
  data_files:
2090
  - split: train
2091
  path: principle_A_domain_3/train-*
2092
+ - config_name: principle_A_reconstruction
2093
+ data_files:
2094
+ - split: train
2095
+ path: principle_A_reconstruction/train-*
2096
  ---
2097
 
2098
  # Dataset Card for "blimp"
dataset_infos.json CHANGED
@@ -3271,62 +3271,50 @@
3271
  "features": {
3272
  "sentence_good": {
3273
  "dtype": "string",
3274
- "id": null,
3275
  "_type": "Value"
3276
  },
3277
  "sentence_bad": {
3278
  "dtype": "string",
3279
- "id": null,
3280
  "_type": "Value"
3281
  },
3282
  "field": {
3283
  "dtype": "string",
3284
- "id": null,
3285
  "_type": "Value"
3286
  },
3287
  "linguistics_term": {
3288
  "dtype": "string",
3289
- "id": null,
3290
  "_type": "Value"
3291
  },
3292
  "UID": {
3293
  "dtype": "string",
3294
- "id": null,
3295
  "_type": "Value"
3296
  },
3297
  "simple_LM_method": {
3298
  "dtype": "bool",
3299
- "id": null,
3300
  "_type": "Value"
3301
  },
3302
  "one_prefix_method": {
3303
  "dtype": "bool",
3304
- "id": null,
3305
  "_type": "Value"
3306
  },
3307
  "two_prefix_method": {
3308
  "dtype": "bool",
3309
- "id": null,
3310
  "_type": "Value"
3311
  },
3312
  "lexically_identical": {
3313
  "dtype": "bool",
3314
- "id": null,
3315
  "_type": "Value"
3316
  },
3317
  "pair_id": {
3318
  "dtype": "int32",
3319
- "id": null,
3320
  "_type": "Value"
3321
  }
3322
  },
3323
- "supervised_keys": null,
3324
  "builder_name": "blimp",
 
3325
  "config_name": "principle_A_reconstruction",
3326
  "version": {
3327
  "version_str": "0.1.0",
3328
- "description": null,
3329
- "datasets_version_to_prepare": null,
3330
  "major": 0,
3331
  "minor": 1,
3332
  "patch": 0
@@ -3334,20 +3322,14 @@
3334
  "splits": {
3335
  "train": {
3336
  "name": "train",
3337
- "num_bytes": 153499,
3338
  "num_examples": 1000,
3339
- "dataset_name": "blimp"
3340
- }
3341
- },
3342
- "download_checksums": {
3343
- "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_reconstruction.jsonl": {
3344
- "num_bytes": 345494,
3345
- "checksum": "65dc403c4d6e9a609f2c4cc2ec6c8baa05cf6fce89b920a816b0d9932204b2a3"
3346
  }
3347
  },
3348
- "download_size": 345494,
3349
- "dataset_size": 153499,
3350
- "size_in_bytes": 498993
3351
  },
3352
  "regular_plural_subject_verb_agreement_1": {
3353
  "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n",
 
3271
  "features": {
3272
  "sentence_good": {
3273
  "dtype": "string",
 
3274
  "_type": "Value"
3275
  },
3276
  "sentence_bad": {
3277
  "dtype": "string",
 
3278
  "_type": "Value"
3279
  },
3280
  "field": {
3281
  "dtype": "string",
 
3282
  "_type": "Value"
3283
  },
3284
  "linguistics_term": {
3285
  "dtype": "string",
 
3286
  "_type": "Value"
3287
  },
3288
  "UID": {
3289
  "dtype": "string",
 
3290
  "_type": "Value"
3291
  },
3292
  "simple_LM_method": {
3293
  "dtype": "bool",
 
3294
  "_type": "Value"
3295
  },
3296
  "one_prefix_method": {
3297
  "dtype": "bool",
 
3298
  "_type": "Value"
3299
  },
3300
  "two_prefix_method": {
3301
  "dtype": "bool",
 
3302
  "_type": "Value"
3303
  },
3304
  "lexically_identical": {
3305
  "dtype": "bool",
 
3306
  "_type": "Value"
3307
  },
3308
  "pair_id": {
3309
  "dtype": "int32",
 
3310
  "_type": "Value"
3311
  }
3312
  },
 
3313
  "builder_name": "blimp",
3314
+ "dataset_name": "blimp",
3315
  "config_name": "principle_A_reconstruction",
3316
  "version": {
3317
  "version_str": "0.1.0",
 
 
3318
  "major": 0,
3319
  "minor": 1,
3320
  "patch": 0
 
3322
  "splits": {
3323
  "train": {
3324
  "name": "train",
3325
+ "num_bytes": 152104,
3326
  "num_examples": 1000,
3327
+ "dataset_name": null
 
 
 
 
 
 
3328
  }
3329
  },
3330
+ "download_size": 44480,
3331
+ "dataset_size": 152104,
3332
+ "size_in_bytes": 196584
3333
  },
3334
  "regular_plural_subject_verb_agreement_1": {
3335
  "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n",
principle_A_reconstruction/train-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd57fef2d8e907730bfdf9be1048bfaa75571b8a5c3481219b08254df544feb1
3
+ size 44480