Datasets:

Languages:
English
ArXiv:
License:
Daniel O'Connell commited on
Commit
1cdb349
·
1 Parent(s): 8964973

update stampy

Browse files
Files changed (3) hide show
  1. README.md +41 -25
  2. aisafety.info.jsonl +0 -0
  3. alignment-research-dataset.py +0 -3
README.md CHANGED
@@ -34,12 +34,6 @@ dataset_info:
34
  dtype: string
35
  - name: summary
36
  sequence: string
37
- - name: question
38
- dtype: string
39
- - name: answer
40
- dtype: string
41
- - name: entry
42
- dtype: string
43
  - name: karma
44
  dtype: int32
45
  - name: votes
@@ -102,10 +96,10 @@ dataset_info:
102
  dtype: string
103
  splits:
104
  - name: train
105
- num_bytes: 319462301
106
- num_examples: 9688
107
- download_size: 366146519
108
- dataset_size: 319462301
109
  - config_name: agentmodels
110
  features:
111
  - name: id
@@ -228,18 +222,12 @@ dataset_info:
228
  sequence: string
229
  - name: source_filetype
230
  dtype: string
231
- - name: question
232
- dtype: string
233
- - name: answer
234
- dtype: string
235
- - name: entry
236
- dtype: string
237
  splits:
238
  - name: train
239
- num_bytes: 1573839
240
- num_examples: 291
241
- download_size: 1616143
242
- dataset_size: 1573839
243
  - config_name: alignmentforum
244
  features:
245
  - name: id
@@ -270,12 +258,14 @@ dataset_info:
270
  dtype: string
271
  - name: source_type
272
  dtype: string
 
 
273
  splits:
274
  - name: train
275
- num_bytes: 73451397
276
  num_examples: 2643
277
  download_size: 75161883
278
- dataset_size: 73451397
279
  - config_name: alignment_newsletter
280
  features:
281
  - name: id
@@ -516,12 +506,14 @@ dataset_info:
516
  sequence:
517
  - name: title
518
  dtype: string
 
 
519
  splits:
520
  - name: train
521
- num_bytes: 2246011
522
  num_examples: 51
523
  download_size: 2435649
524
- dataset_size: 2246011
525
  - config_name: eaforum
526
  features:
527
  - name: id
@@ -966,6 +958,30 @@ dataset_info:
966
  num_examples: 23
967
  download_size: 604224
968
  dataset_size: 594207
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
969
  ---
970
  # AI Alignment Research Dataset
971
 
@@ -1055,4 +1071,4 @@ This README contains info about the number of rows and their features which shou
1055
 
1056
  For more information, here is the [paper](https://arxiv.org/abs/2206.02841) and [LessWrong](https://www.lesswrong.com/posts/FgjcHiWvADgsocE34/a-descriptive-not-prescriptive-overview-of-current-ai) post. Please use the following citation when using the dataset:
1057
 
1058
- Kirchner, J. H., Smith, L., Thibodeau, J., McDonnell, K., and Reynolds, L. "Understanding AI alignment research: A Systematic Analysis." arXiv preprint arXiv:2022.4338861 (2022).
 
34
  dtype: string
35
  - name: summary
36
  sequence: string
 
 
 
 
 
 
37
  - name: karma
38
  dtype: int32
39
  - name: votes
 
96
  dtype: string
97
  splits:
98
  - name: train
99
+ num_bytes: 318682720
100
+ num_examples: 9700
101
+ download_size: 365472655
102
+ dataset_size: 318682720
103
  - config_name: agentmodels
104
  features:
105
  - name: id
 
222
  sequence: string
223
  - name: source_filetype
224
  dtype: string
 
 
 
 
 
 
225
  splits:
226
  - name: train
227
+ num_bytes: 798694
228
+ num_examples: 294
229
+ download_size: 830259
230
+ dataset_size: 798694
231
  - config_name: alignmentforum
232
  features:
233
  - name: id
 
258
  dtype: string
259
  - name: source_type
260
  dtype: string
261
+ - name: summary
262
+ sequence: string
263
  splits:
264
  - name: train
265
+ num_bytes: 73852390
266
  num_examples: 2643
267
  download_size: 75161883
268
+ dataset_size: 73852390
269
  - config_name: alignment_newsletter
270
  features:
271
  - name: id
 
506
  sequence:
507
  - name: title
508
  dtype: string
509
+ - name: summary
510
+ sequence: string
511
  splits:
512
  - name: train
513
+ num_bytes: 2251795
514
  num_examples: 51
515
  download_size: 2435649
516
+ dataset_size: 2251795
517
  - config_name: eaforum
518
  features:
519
  - name: id
 
958
  num_examples: 23
959
  download_size: 604224
960
  dataset_size: 594207
961
+ - config_name: ml_safety_newsletter
962
+ features:
963
+ - name: id
964
+ dtype: string
965
+ - name: source
966
+ dtype: string
967
+ - name: title
968
+ dtype: string
969
+ - name: text
970
+ dtype: large_string
971
+ - name: url
972
+ dtype: string
973
+ - name: date_published
974
+ dtype: string
975
+ - name: authors
976
+ sequence: string
977
+ - name: source_type
978
+ dtype: string
979
+ splits:
980
+ - name: train
981
+ num_bytes: 110266
982
+ num_examples: 9
983
+ download_size: 112020
984
+ dataset_size: 110266
985
  ---
986
  # AI Alignment Research Dataset
987
 
 
1071
 
1072
  For more information, here is the [paper](https://arxiv.org/abs/2206.02841) and [LessWrong](https://www.lesswrong.com/posts/FgjcHiWvADgsocE34/a-descriptive-not-prescriptive-overview-of-current-ai) post. Please use the following citation when using the dataset:
1073
 
1074
+ Kirchner, J. H., Smith, L., Thibodeau, J., McDonnell, K., and Reynolds, L. "Understanding AI alignment research: A Systematic Analysis." arXiv preprint arXiv:2022.4338861 (2022).
aisafety.info.jsonl CHANGED
The diff for this file is too large to render. See raw diff
 
alignment-research-dataset.py CHANGED
@@ -113,9 +113,6 @@ DATASOURCES = {
113
  },
114
  'aisafety.info': {
115
  'source_filetype': Value(dtype='string'),
116
- 'question': Value(dtype='string'),
117
- 'answer': Value(dtype='string'),
118
- 'entry': Value(dtype='string')
119
  },
120
  'alignmentforum': {
121
  'karma': Value(dtype='int32'),
 
113
  },
114
  'aisafety.info': {
115
  'source_filetype': Value(dtype='string'),
 
 
 
116
  },
117
  'alignmentforum': {
118
  'karma': Value(dtype='int32'),