Daniel O'Connell
commited on
Commit
·
1cdb349
1
Parent(s):
8964973
update stampy
Browse files- README.md +41 -25
- aisafety.info.jsonl +0 -0
- 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:
|
106 |
-
num_examples:
|
107 |
-
download_size:
|
108 |
-
dataset_size:
|
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:
|
240 |
-
num_examples:
|
241 |
-
download_size:
|
242 |
-
dataset_size:
|
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:
|
276 |
num_examples: 2643
|
277 |
download_size: 75161883
|
278 |
-
dataset_size:
|
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:
|
522 |
num_examples: 51
|
523 |
download_size: 2435649
|
524 |
-
dataset_size:
|
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'),
|