sileod commited on
Commit
1a22a9e
·
verified ·
1 Parent(s): 8fea838

Add tasksource-jev-typed-decisions dataset card

Browse files
Files changed (1) hide show
  1. README.md +1 -103
README.md CHANGED
@@ -19,108 +19,6 @@ tags:
19
  - decision-model
20
  size_categories:
21
  - 1M<n<10M
22
- configs:
23
- - config_name: default
24
- data_files:
25
- - split: train
26
- path: data/train-*
27
- - split: validation
28
- path: data/validation-*
29
- - split: test
30
- path: data/test-*
31
- - config_name: full
32
- data_files:
33
- - split: train
34
- path: full/train-*
35
- - split: validation
36
- path: full/validation-*
37
- - split: test
38
- path: full/test-*
39
- dataset_info:
40
- - config_name: default
41
- features:
42
- - name: state
43
- dtype: string
44
- - name: kind
45
- dtype: string
46
- - name: id
47
- dtype: string
48
- - name: options
49
- list: string
50
- - name: target
51
- list: float64
52
- - name: question
53
- dtype: string
54
- - name: source
55
- dtype: string
56
- - name: variant
57
- dtype: string
58
- - name: split
59
- dtype: string
60
- - name: group_id
61
- dtype: string
62
- - name: question_id
63
- dtype: string
64
- - name: example_id
65
- dtype: string
66
- - name: license
67
- dtype: string
68
- - name: license_use
69
- dtype: string
70
- splits:
71
- - name: train
72
- num_bytes: 1539759841
73
- num_examples: 1006875
74
- - name: validation
75
- num_bytes: 25864050
76
- num_examples: 14620
77
- - name: test
78
- num_bytes: 22502046
79
- num_examples: 14829
80
- download_size: 669264666
81
- dataset_size: 1588125937
82
- - config_name: full
83
- features:
84
- - name: state
85
- dtype: string
86
- - name: kind
87
- dtype: string
88
- - name: id
89
- dtype: string
90
- - name: options
91
- list: string
92
- - name: target
93
- list: float64
94
- - name: question
95
- dtype: string
96
- - name: source
97
- dtype: string
98
- - name: variant
99
- dtype: string
100
- - name: split
101
- dtype: string
102
- - name: group_id
103
- dtype: string
104
- - name: question_id
105
- dtype: string
106
- - name: example_id
107
- dtype: string
108
- - name: license
109
- dtype: string
110
- - name: license_use
111
- dtype: string
112
- splits:
113
- - name: train
114
- num_bytes: 3825958901
115
- num_examples: 2507000
116
- - name: validation
117
- num_bytes: 27795181
118
- num_examples: 15700
119
- - name: test
120
- num_bytes: 23831256
121
- num_examples: 15700
122
- download_size: 1837838436
123
- dataset_size: 3877585338
124
  ---
125
 
126
  # tasksource-jev-typed-decisions
@@ -176,7 +74,7 @@ Rows sharing `group_id` can be assembled into a multi-question request. Vision i
176
 
177
  ## Construction and evaluation
178
 
179
- Tasksource standardizes native labels and choices, adds task-specific questions where needed, and applies deterministic option permutations and low-frequency text variants. Some text rows pack related decisions over a shared state. A decision's majority label keeps at most three rows per row of its other labels (never fewer than 250; graded answers count as yes above 0.5), so skewed binary sources such as answer selection or toxicity no longer reward always answering the common label; procedural questions keep each label under 60%. The visual config contains direct decisions and shares sampling budgets across related task views.
180
 
181
  Native source partitions are retained where available; derived group-based holdouts and overlap exclusions are documented per source. These checks do not establish absence of all cross-dataset or benchmark contamination. For example, Mind2Web text tasks use derived holdouts from public training trajectories, not official benchmark tests; alignment with private Decision Index evaluation rows is unverified. Mind2Web is excluded from the visual config.
182
 
 
19
  - decision-model
20
  size_categories:
21
  - 1M<n<10M
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
22
  ---
23
 
24
  # tasksource-jev-typed-decisions
 
74
 
75
  ## Construction and evaluation
76
 
77
+ Tasksource standardizes native labels and choices, adds task-specific questions where needed, and applies deterministic option permutations and low-frequency text variants. Some text rows pack related decisions over a shared state. A decision's majority label keeps at most three rows per row of its other labels (graded answers count as yes above 0.5; a question with a single label is dropped), so skewed binary sources such as answer selection or toxicity no longer reward always answering the common label; procedural questions keep each label under 60%. The visual config contains direct decisions and shares sampling budgets across related task views.
78
 
79
  Native source partitions are retained where available; derived group-based holdouts and overlap exclusions are documented per source. These checks do not establish absence of all cross-dataset or benchmark contamination. For example, Mind2Web text tasks use derived holdouts from public training trajectories, not official benchmark tests; alignment with private Decision Index evaluation rows is unverified. Mind2Web is excluded from the visual config.
80