File size: 3,156 Bytes
36d3276
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
operators: 
- id: load_math_alg
  config:
    type: hf_source
    dataset: deepmind/math_dataset
    split: train
    subset: algebra__linear_2d_composed
    trust_remote_code: True
- id: load_math_poly_eval
  config:
    type: hf_source
    dataset: deepmind/math_dataset
    split: train
    subset: polynomials__evaluate_composed
    trust_remote_code: True
- id: load_math_prob_seq
  config:
    type: hf_source
    dataset: deepmind/math_dataset
    split: train
    subset: probability__swr_p_sequence
    trust_remote_code: True
- id: load_math_prob_level
  config:
    type: hf_source
    dataset: deepmind/math_dataset
    split: train
    subset: probability__swr_p_level_set
    trust_remote_code: True
- id: load_math_poly_pow
  config:
    type: hf_source
    dataset: deepmind/math_dataset
    split: train
    subset: polynomials__simplify_power
    trust_remote_code: True
- id: load_math_calc
  config:
    type: hf_source
    dataset: deepmind/math_dataset
    split: train
    subset: calculus__differentiate_composed
    trust_remote_code: True
- id: clean_up
  config:
    type: function
    function: data_strategies.SeedMathData.utils.deepmind_map
  input_ids:
  - load_math_calc
  - load_math_poly_eval
  - load_math_alg
  - load_math_prob_seq
  - load_math_prob_level
  - load_math_poly_pow
- id: sample_dataset_pre
  config:
    type: function
    function: data_strategies.commons.uniform_sample_fixed
    function_config:
      num_samples: 1_100_000
  input_ids:
  - clean_up

- id: add_source_name
  config:
    type: function
    function: data_strategies.commons.add_constant_columns
    function_config:
      _source: "deepmind/math_dataset" 
  input_ids:
  - sample_dataset_pre
- id: decontaminate
  config:
    type: cpu_function
    sharded: true
    num_shards: 8
    num_cpus: 32
    function:  data_strategies._A1_Source.utils.decontaminate_fuzzy_w_ngram
    function_config:
      column: instruction_seed
      eval_datasets:
        - HuggingFaceH4/MATH-500
        - Maxwell-Jia/AIME_2024
        - AI-MO/aimo-validation-amc
        - livecodebench/code_generation_lite
        - mlfoundations-dev/AIME2025_combined
        - cais/hle
        - open-r1/codeforces
        - Idavidrein/gpqa
        - daman1209arora/jeebench
        - mlfoundations-dev/mmlu_pro_eval_full
        - Qwen/CodeElo
        - open-r1/ioi
      eval_columns:
        - problem
        - Problem
        - problem
        - question_content
        - question
        - question
        - description
        - Question
        - question
        - prompt
        - description
        - statement
      eval_splits:
        - test
        - train
        - train
        - test
        - train
        - test
        - test
        - train
        - test
        - test
        - test
        - test
      eval_subsets:
        Idavidrein/gpqa: gpqa_diamond
      similarity_threshold: 75.0
      ngram_size: 13
  input_ids:
  - add_source_name
- id: sample_dataset
  config:
    type: function
    function: data_strategies.commons.uniform_sample_fixed
    function_config:
      num_samples: 1_000_000
  input_ids: ["decontaminate"]