ClaasBeger commited on
Commit
3346cb5
·
verified ·
1 Parent(s): 80c589a

Upload ConceptARC_Rule_Annotations_croissant_merged.json

Browse files
ConceptARC_Rule_Annotations_croissant_merged.json ADDED
@@ -0,0 +1,687 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "@context": {
3
+ "@language": "en",
4
+ "@vocab": "https://schema.org/",
5
+ "arrayShape": "cr:arrayShape",
6
+ "citeAs": "cr:citeAs",
7
+ "column": "cr:column",
8
+ "conformsTo": "dct:conformsTo",
9
+ "containedIn": "cr:containedIn",
10
+ "cr": "http://mlcommons.org/croissant/",
11
+ "data": {
12
+ "@id": "cr:data",
13
+ "@type": "@json"
14
+ },
15
+ "dataBiases": "cr:dataBiases",
16
+ "dataCollection": "cr:dataCollection",
17
+ "dataType": {
18
+ "@id": "cr:dataType",
19
+ "@type": "@vocab"
20
+ },
21
+ "dct": "http://purl.org/dc/terms/",
22
+ "extract": "cr:extract",
23
+ "field": "cr:field",
24
+ "fileProperty": "cr:fileProperty",
25
+ "fileObject": "cr:fileObject",
26
+ "fileSet": "cr:fileSet",
27
+ "format": "cr:format",
28
+ "includes": "cr:includes",
29
+ "isArray": "cr:isArray",
30
+ "isLiveDataset": "cr:isLiveDataset",
31
+ "jsonPath": "cr:jsonPath",
32
+ "key": "cr:key",
33
+ "md5": "cr:md5",
34
+ "parentField": "cr:parentField",
35
+ "path": "cr:path",
36
+ "personalSensitiveInformation": "cr:personalSensitiveInformation",
37
+ "recordSet": "cr:recordSet",
38
+ "references": "cr:references",
39
+ "regex": "cr:regex",
40
+ "repeated": "cr:repeated",
41
+ "replace": "cr:replace",
42
+ "sc": "https://schema.org/",
43
+ "separator": "cr:separator",
44
+ "source": "cr:source",
45
+ "subField": "cr:subField",
46
+ "transform": "cr:transform",
47
+ "rai": "http://mlcommons.org/croissant/RAI/",
48
+ "prov": "http://www.w3.org/ns/prov#"
49
+ },
50
+ "@type": "sc:Dataset",
51
+ "distribution": [
52
+ {
53
+ "@type": "cr:FileObject",
54
+ "@id": "repo",
55
+ "name": "repo",
56
+ "description": "The Hugging Face git repository.",
57
+ "contentUrl": "https://huggingface.co/datasets/ClaasBeger/ConceptARC_Rule_Annotations/tree/refs%2Fconvert%2Fparquet",
58
+ "encodingFormat": "git+https",
59
+ "sha256": "https://github.com/mlcommons/croissant/issues/80"
60
+ },
61
+ {
62
+ "@type": "cr:FileSet",
63
+ "@id": "parquet-files-for-config-model_evaluations",
64
+ "containedIn": {
65
+ "@id": "repo"
66
+ },
67
+ "encodingFormat": "application/x-parquet",
68
+ "includes": "model_evaluations/*/*.parquet"
69
+ },
70
+ {
71
+ "@type": "cr:FileSet",
72
+ "@id": "parquet-files-for-config-ground_truth_rules",
73
+ "containedIn": {
74
+ "@id": "repo"
75
+ },
76
+ "encodingFormat": "application/x-parquet",
77
+ "includes": "ground_truth_rules/*/*.parquet"
78
+ },
79
+ {
80
+ "@type": "cr:FileSet",
81
+ "@id": "parquet-files-for-config-human_rule_annotations",
82
+ "containedIn": {
83
+ "@id": "repo"
84
+ },
85
+ "encodingFormat": "application/x-parquet",
86
+ "includes": "human_rule_annotations/*/*.parquet"
87
+ }
88
+ ],
89
+ "recordSet": [
90
+ {
91
+ "@type": "cr:RecordSet",
92
+ "dataType": "cr:Split",
93
+ "key": {
94
+ "@id": "model_evaluations_splits/split_name"
95
+ },
96
+ "@id": "model_evaluations_splits",
97
+ "name": "model_evaluations_splits",
98
+ "description": "Splits for the model_evaluations config.",
99
+ "field": [
100
+ {
101
+ "@type": "cr:Field",
102
+ "@id": "model_evaluations_splits/split_name",
103
+ "dataType": "sc:Text"
104
+ }
105
+ ],
106
+ "data": [
107
+ {
108
+ "model_evaluations_splits/split_name": "train"
109
+ }
110
+ ]
111
+ },
112
+ {
113
+ "@type": "cr:RecordSet",
114
+ "@id": "model_evaluations",
115
+ "description": "ClaasBeger/ConceptARC_Rule_Annotations - 'model_evaluations' subset",
116
+ "field": [
117
+ {
118
+ "@type": "cr:Field",
119
+ "@id": "model_evaluations/split",
120
+ "dataType": "sc:Text",
121
+ "source": {
122
+ "fileSet": {
123
+ "@id": "parquet-files-for-config-model_evaluations"
124
+ },
125
+ "extract": {
126
+ "fileProperty": "fullpath"
127
+ },
128
+ "transform": {
129
+ "regex": "model_evaluations/(?:partial-)?(train)/.+parquet$"
130
+ }
131
+ },
132
+ "references": {
133
+ "field": {
134
+ "@id": "model_evaluations_splits/split_name"
135
+ }
136
+ }
137
+ },
138
+ {
139
+ "@type": "cr:Field",
140
+ "@id": "model_evaluations/modality",
141
+ "dataType": "sc:Text",
142
+ "source": {
143
+ "fileSet": {
144
+ "@id": "parquet-files-for-config-model_evaluations"
145
+ },
146
+ "extract": {
147
+ "column": "modality"
148
+ }
149
+ }
150
+ },
151
+ {
152
+ "@type": "cr:Field",
153
+ "@id": "model_evaluations/model",
154
+ "dataType": "sc:Text",
155
+ "source": {
156
+ "fileSet": {
157
+ "@id": "parquet-files-for-config-model_evaluations"
158
+ },
159
+ "extract": {
160
+ "column": "model"
161
+ }
162
+ }
163
+ },
164
+ {
165
+ "@type": "cr:Field",
166
+ "@id": "model_evaluations/run_folder",
167
+ "dataType": "sc:Text",
168
+ "source": {
169
+ "fileSet": {
170
+ "@id": "parquet-files-for-config-model_evaluations"
171
+ },
172
+ "extract": {
173
+ "column": "run_folder"
174
+ }
175
+ }
176
+ },
177
+ {
178
+ "@type": "cr:Field",
179
+ "@id": "model_evaluations/source_file",
180
+ "dataType": "sc:Text",
181
+ "source": {
182
+ "fileSet": {
183
+ "@id": "parquet-files-for-config-model_evaluations"
184
+ },
185
+ "extract": {
186
+ "column": "source_file"
187
+ }
188
+ }
189
+ },
190
+ {
191
+ "@type": "cr:Field",
192
+ "@id": "model_evaluations/timestamp",
193
+ "dataType": "sc:Text",
194
+ "source": {
195
+ "fileSet": {
196
+ "@id": "parquet-files-for-config-model_evaluations"
197
+ },
198
+ "extract": {
199
+ "column": "timestamp"
200
+ }
201
+ }
202
+ },
203
+ {
204
+ "@type": "cr:Field",
205
+ "@id": "model_evaluations/concept",
206
+ "dataType": "sc:Text",
207
+ "source": {
208
+ "fileSet": {
209
+ "@id": "parquet-files-for-config-model_evaluations"
210
+ },
211
+ "extract": {
212
+ "column": "concept"
213
+ }
214
+ }
215
+ },
216
+ {
217
+ "@type": "cr:Field",
218
+ "@id": "model_evaluations/puzzle",
219
+ "dataType": "sc:Text",
220
+ "source": {
221
+ "fileSet": {
222
+ "@id": "parquet-files-for-config-model_evaluations"
223
+ },
224
+ "extract": {
225
+ "column": "puzzle"
226
+ }
227
+ }
228
+ },
229
+ {
230
+ "@type": "cr:Field",
231
+ "@id": "model_evaluations/test_idx",
232
+ "dataType": "cr:Int64",
233
+ "source": {
234
+ "fileSet": {
235
+ "@id": "parquet-files-for-config-model_evaluations"
236
+ },
237
+ "extract": {
238
+ "column": "test_idx"
239
+ }
240
+ }
241
+ },
242
+ {
243
+ "@type": "cr:Field",
244
+ "@id": "model_evaluations/effort",
245
+ "dataType": "sc:Text",
246
+ "source": {
247
+ "fileSet": {
248
+ "@id": "parquet-files-for-config-model_evaluations"
249
+ },
250
+ "extract": {
251
+ "column": "effort"
252
+ }
253
+ }
254
+ },
255
+ {
256
+ "@type": "cr:Field",
257
+ "@id": "model_evaluations/is_correct",
258
+ "dataType": "cr:Int64",
259
+ "source": {
260
+ "fileSet": {
261
+ "@id": "parquet-files-for-config-model_evaluations"
262
+ },
263
+ "extract": {
264
+ "column": "is_correct"
265
+ }
266
+ }
267
+ },
268
+ {
269
+ "@type": "cr:Field",
270
+ "@id": "model_evaluations/err",
271
+ "dataType": "sc:Text",
272
+ "source": {
273
+ "fileSet": {
274
+ "@id": "parquet-files-for-config-model_evaluations"
275
+ },
276
+ "extract": {
277
+ "column": "err"
278
+ }
279
+ }
280
+ },
281
+ {
282
+ "@type": "cr:Field",
283
+ "@id": "model_evaluations/tools_used",
284
+ "dataType": "cr:Int64",
285
+ "source": {
286
+ "fileSet": {
287
+ "@id": "parquet-files-for-config-model_evaluations"
288
+ },
289
+ "extract": {
290
+ "column": "tools_used"
291
+ }
292
+ }
293
+ },
294
+ {
295
+ "@type": "cr:Field",
296
+ "@id": "model_evaluations/answer",
297
+ "dataType": "sc:Text",
298
+ "source": {
299
+ "fileSet": {
300
+ "@id": "parquet-files-for-config-model_evaluations"
301
+ },
302
+ "extract": {
303
+ "column": "answer"
304
+ }
305
+ }
306
+ },
307
+ {
308
+ "@type": "cr:Field",
309
+ "@id": "model_evaluations/summary",
310
+ "dataType": "sc:Text",
311
+ "source": {
312
+ "fileSet": {
313
+ "@id": "parquet-files-for-config-model_evaluations"
314
+ },
315
+ "extract": {
316
+ "column": "summary"
317
+ }
318
+ }
319
+ },
320
+ {
321
+ "@type": "cr:Field",
322
+ "@id": "model_evaluations/programming_calls",
323
+ "dataType": "sc:Text",
324
+ "source": {
325
+ "fileSet": {
326
+ "@id": "parquet-files-for-config-model_evaluations"
327
+ },
328
+ "extract": {
329
+ "column": "programming_calls"
330
+ }
331
+ }
332
+ },
333
+ {
334
+ "@type": "cr:Field",
335
+ "@id": "model_evaluations/Rule",
336
+ "dataType": "sc:Text",
337
+ "source": {
338
+ "fileSet": {
339
+ "@id": "parquet-files-for-config-model_evaluations"
340
+ },
341
+ "extract": {
342
+ "column": "Rule"
343
+ }
344
+ }
345
+ },
346
+ {
347
+ "@type": "cr:Field",
348
+ "@id": "model_evaluations/Rule_correct",
349
+ "dataType": "cr:Int64",
350
+ "source": {
351
+ "fileSet": {
352
+ "@id": "parquet-files-for-config-model_evaluations"
353
+ },
354
+ "extract": {
355
+ "column": "Rule_correct"
356
+ }
357
+ }
358
+ },
359
+ {
360
+ "@type": "cr:Field",
361
+ "@id": "model_evaluations/rule_evaluation_reasoning",
362
+ "dataType": "sc:Text",
363
+ "source": {
364
+ "fileSet": {
365
+ "@id": "parquet-files-for-config-model_evaluations"
366
+ },
367
+ "extract": {
368
+ "column": "rule_evaluation_reasoning"
369
+ }
370
+ }
371
+ },
372
+ {
373
+ "@type": "cr:Field",
374
+ "@id": "model_evaluations/Rule_correct_label",
375
+ "dataType": "sc:Text",
376
+ "source": {
377
+ "fileSet": {
378
+ "@id": "parquet-files-for-config-model_evaluations"
379
+ },
380
+ "extract": {
381
+ "column": "Rule_correct_label"
382
+ }
383
+ }
384
+ },
385
+ {
386
+ "@type": "cr:Field",
387
+ "@id": "model_evaluations/starred",
388
+ "dataType": "cr:Int64",
389
+ "source": {
390
+ "fileSet": {
391
+ "@id": "parquet-files-for-config-model_evaluations"
392
+ },
393
+ "extract": {
394
+ "column": "starred"
395
+ }
396
+ }
397
+ }
398
+ ]
399
+ },
400
+ {
401
+ "@type": "cr:RecordSet",
402
+ "dataType": "cr:Split",
403
+ "key": {
404
+ "@id": "ground_truth_rules_splits/split_name"
405
+ },
406
+ "@id": "ground_truth_rules_splits",
407
+ "name": "ground_truth_rules_splits",
408
+ "description": "Splits for the ground_truth_rules config.",
409
+ "field": [
410
+ {
411
+ "@type": "cr:Field",
412
+ "@id": "ground_truth_rules_splits/split_name",
413
+ "dataType": "sc:Text"
414
+ }
415
+ ],
416
+ "data": [
417
+ {
418
+ "ground_truth_rules_splits/split_name": "train"
419
+ }
420
+ ]
421
+ },
422
+ {
423
+ "@type": "cr:RecordSet",
424
+ "@id": "ground_truth_rules",
425
+ "description": "ClaasBeger/ConceptARC_Rule_Annotations - 'ground_truth_rules' subset",
426
+ "field": [
427
+ {
428
+ "@type": "cr:Field",
429
+ "@id": "ground_truth_rules/split",
430
+ "dataType": "sc:Text",
431
+ "source": {
432
+ "fileSet": {
433
+ "@id": "parquet-files-for-config-ground_truth_rules"
434
+ },
435
+ "extract": {
436
+ "fileProperty": "fullpath"
437
+ },
438
+ "transform": {
439
+ "regex": "ground_truth_rules/(?:partial-)?(train)/.+parquet$"
440
+ }
441
+ },
442
+ "references": {
443
+ "field": {
444
+ "@id": "ground_truth_rules_splits/split_name"
445
+ }
446
+ }
447
+ },
448
+ {
449
+ "@type": "cr:Field",
450
+ "@id": "ground_truth_rules/concept",
451
+ "dataType": "sc:Text",
452
+ "source": {
453
+ "fileSet": {
454
+ "@id": "parquet-files-for-config-ground_truth_rules"
455
+ },
456
+ "extract": {
457
+ "column": "concept"
458
+ }
459
+ }
460
+ },
461
+ {
462
+ "@type": "cr:Field",
463
+ "@id": "ground_truth_rules/puzzle",
464
+ "dataType": "sc:Text",
465
+ "source": {
466
+ "fileSet": {
467
+ "@id": "parquet-files-for-config-ground_truth_rules"
468
+ },
469
+ "extract": {
470
+ "column": "puzzle"
471
+ }
472
+ }
473
+ },
474
+ {
475
+ "@type": "cr:Field",
476
+ "@id": "ground_truth_rules/rule",
477
+ "dataType": "sc:Text",
478
+ "source": {
479
+ "fileSet": {
480
+ "@id": "parquet-files-for-config-ground_truth_rules"
481
+ },
482
+ "extract": {
483
+ "column": "rule"
484
+ }
485
+ }
486
+ }
487
+ ]
488
+ },
489
+ {
490
+ "@type": "cr:RecordSet",
491
+ "dataType": "cr:Split",
492
+ "key": {
493
+ "@id": "human_rule_annotations_splits/split_name"
494
+ },
495
+ "@id": "human_rule_annotations_splits",
496
+ "name": "human_rule_annotations_splits",
497
+ "description": "Splits for the human_rule_annotations config.",
498
+ "field": [
499
+ {
500
+ "@type": "cr:Field",
501
+ "@id": "human_rule_annotations_splits/split_name",
502
+ "dataType": "sc:Text"
503
+ }
504
+ ],
505
+ "data": [
506
+ {
507
+ "human_rule_annotations_splits/split_name": "train"
508
+ }
509
+ ]
510
+ },
511
+ {
512
+ "@type": "cr:RecordSet",
513
+ "@id": "human_rule_annotations",
514
+ "description": "ClaasBeger/ConceptARC_Rule_Annotations - 'human_rule_annotations' subset",
515
+ "field": [
516
+ {
517
+ "@type": "cr:Field",
518
+ "@id": "human_rule_annotations/split",
519
+ "dataType": "sc:Text",
520
+ "source": {
521
+ "fileSet": {
522
+ "@id": "parquet-files-for-config-human_rule_annotations"
523
+ },
524
+ "extract": {
525
+ "fileProperty": "fullpath"
526
+ },
527
+ "transform": {
528
+ "regex": "human_rule_annotations/(?:partial-)?(train)/.+parquet$"
529
+ }
530
+ },
531
+ "references": {
532
+ "field": {
533
+ "@id": "human_rule_annotations_splits/split_name"
534
+ }
535
+ }
536
+ },
537
+ {
538
+ "@type": "cr:Field",
539
+ "@id": "human_rule_annotations/concept_group",
540
+ "dataType": "sc:Text",
541
+ "source": {
542
+ "fileSet": {
543
+ "@id": "parquet-files-for-config-human_rule_annotations"
544
+ },
545
+ "extract": {
546
+ "column": "concept_group"
547
+ }
548
+ }
549
+ },
550
+ {
551
+ "@type": "cr:Field",
552
+ "@id": "human_rule_annotations/Task",
553
+ "dataType": "sc:Text",
554
+ "source": {
555
+ "fileSet": {
556
+ "@id": "parquet-files-for-config-human_rule_annotations"
557
+ },
558
+ "extract": {
559
+ "column": "Task"
560
+ }
561
+ }
562
+ },
563
+ {
564
+ "@type": "cr:Field",
565
+ "@id": "human_rule_annotations/Test",
566
+ "dataType": "cr:Int64",
567
+ "source": {
568
+ "fileSet": {
569
+ "@id": "parquet-files-for-config-human_rule_annotations"
570
+ },
571
+ "extract": {
572
+ "column": "Test"
573
+ }
574
+ }
575
+ },
576
+ {
577
+ "@type": "cr:Field",
578
+ "@id": "human_rule_annotations/VerbalDescription",
579
+ "dataType": "sc:Text",
580
+ "source": {
581
+ "fileSet": {
582
+ "@id": "parquet-files-for-config-human_rule_annotations"
583
+ },
584
+ "extract": {
585
+ "column": "VerbalDescription"
586
+ }
587
+ }
588
+ },
589
+ {
590
+ "@type": "cr:Field",
591
+ "@id": "human_rule_annotations/Rule_correct",
592
+ "dataType": "cr:Int64",
593
+ "source": {
594
+ "fileSet": {
595
+ "@id": "parquet-files-for-config-human_rule_annotations"
596
+ },
597
+ "extract": {
598
+ "column": "Rule_correct"
599
+ }
600
+ }
601
+ },
602
+ {
603
+ "@type": "cr:Field",
604
+ "@id": "human_rule_annotations/Rule_correct_label",
605
+ "dataType": "sc:Text",
606
+ "source": {
607
+ "fileSet": {
608
+ "@id": "parquet-files-for-config-human_rule_annotations"
609
+ },
610
+ "extract": {
611
+ "column": "Rule_correct_label"
612
+ }
613
+ }
614
+ },
615
+ {
616
+ "@type": "cr:Field",
617
+ "@id": "human_rule_annotations/starred",
618
+ "dataType": "cr:Int64",
619
+ "source": {
620
+ "fileSet": {
621
+ "@id": "parquet-files-for-config-human_rule_annotations"
622
+ },
623
+ "extract": {
624
+ "column": "starred"
625
+ }
626
+ }
627
+ }
628
+ ]
629
+ }
630
+ ],
631
+ "conformsTo": "http://mlcommons.org/croissant/1.1",
632
+ "name": "ConceptARC_Rule_Annotations",
633
+ "description": "\n\t\n\t\t\n\t\n\t\n\t\tConceptARC Rule Annotations\n\t\n\nModel outputs, natural-language rules and human judgements of those rules on the 480 tasks of the ConceptARC benchmark. This is the data behind the paper Distinguishing Performance From Competence in Evaluations of Humanlike Abstract Reasoning (NeurIPS 2026, Evaluations and Datasets Track).\n\nPaper: arXiv:2510.02125\nProject page and interactive viewer: claasbeger.github.io/performance-competence-gap\nAuthors: Claas Beger, Ryan Yi, Shuhao Fu, Kaleda… See the full description on the dataset page: https://huggingface.co/datasets/ClaasBeger/ConceptARC_Rule_Annotations.",
634
+ "alternateName": [
635
+ "ClaasBeger/ConceptARC_Rule_Annotations",
636
+ "ConceptARC Rule Annotations"
637
+ ],
638
+ "creator": {
639
+ "@type": "Person",
640
+ "name": "Claas Beger",
641
+ "url": "https://huggingface.co/ClaasBeger"
642
+ },
643
+ "keywords": [
644
+ "English",
645
+ "mit",
646
+ "10K - 100K",
647
+ "parquet",
648
+ "Tabular",
649
+ "Text",
650
+ "Datasets",
651
+ "pandas",
652
+ "Polars",
653
+ "Croissant",
654
+ "arxiv:2510.02125",
655
+ "🇺🇸 Region: US",
656
+ "arc",
657
+ "concept-arc",
658
+ "abstract-reasoning",
659
+ "rule-learning",
660
+ "evaluation"
661
+ ],
662
+ "license": "https://choosealicense.com/licenses/mit/",
663
+ "url": "https://huggingface.co/datasets/ClaasBeger/ConceptARC_Rule_Annotations",
664
+ "rai:dataLimitations": "This dataset is limited to evaluation based on three categories including Correct-Intended, Correct-Unintended and Incorrect Rules. Results have been reached through a consensus of multiple annotators, but may still reflect subjectivity to a certain degree. In addition, the annotations do not usually consider the entire reasoning summary exhibited by the different models.",
665
+ "rai:dataBiases": "The dataset re-uses human rules sourced from participants through the online platforms Prolific and Amazon Mechanical Turk, inheriting the distribution described in “The ConceptARC Benchmark: Evaluating Understanding and Generalization in the ARC Domain” (Moskvichev et al., 2023). The present work uses unpublished data provided by the authors of that study.",
666
+ "rai:personalSensitiveInformation": "This dataset does not contain sensitive private information.",
667
+ "rai:dataUseCases": "This dataset evaluates and compares the ability to produce human-like natural language solutions to Abstract Reasoning tasks of the ConceptARC benchmark. We do not claim validation on real-world use cases.",
668
+ "rai:dataSocialImpact": "Constitutes a meaningful first instantiation of human-AI comparison on abstract reasoning tasks. We do not anticipate significant societal effects of this dataset.",
669
+ "rai:hasSyntheticData": false,
670
+ "prov:wasDerivedFrom": [
671
+ {
672
+ "@id": "https://github.com/victorvikram/conceptarc",
673
+ "prov:label": "ConceptARC",
674
+ "sc:license": "MIT"
675
+ }
676
+ ],
677
+ "prov:wasGeneratedBy": [
678
+ {
679
+ "@type": "prov:Activity",
680
+ "prov:type": {
681
+ "@id": "https://www.wikidata.org/wiki/Q109719325"
682
+ },
683
+ "prov:label": "Rule Evaluation",
684
+ "sc:description": "Evaluation of alignment of proposed natural language rules to solve visual abstract reasoning tasks using three distinct categories (\"Correct-Intended\", \"Correct-Unintended\", \"Incorrect\"). We use a standard protocol of independent annotations by two members of our team, with a following discussion for any disagreements."
685
+ }
686
+ ]
687
+ }