Dataset Viewer
Auto-converted to Parquet Duplicate
ctx
stringlengths
7
26
answer
stringclasses
1k values
30 + 8 * (21 + 45) =
558
23 + 22 + 25 - 13 =
57
20 + 26 =
46
61 - 7 + 19 =
73
11 * 13 =
143
(24 + 20) - 8 =
36
5 + 13 * 49 =
642
17 * 58 - 42 - 69 =
875
11 * 18 =
198
79 + (9 - 13) * 10 =
39
63 * 14 + 53 =
935
16 + 102 =
118
27 + 20 * 29 - 25 =
582
4 + 10 + (11 * 19) =
223
15 - 6 =
9
26 + 13 =
39
(115 + 25) - (12 - 20) =
148
37 - 19 =
18
60 + 66 + 4 =
130
57 + 5 =
62
17 - 106 + 61 + 49 =
21
10 * 2 =
20
9 * 13 =
117
77 + 19 * 19 =
438
80 - 19 =
61
(10 * 11) + 5 =
115
(11 + 28) + 0 * 21 =
39
(57 + 7) - 5 + 13 =
72
17 * 13 =
221
(13 * 10) - 11 =
119
23 + 112 =
135
15 * 2 =
30
(24 + 20) + (30 * 21) =
674
25 + 1 - (8 - 10) =
28
20 - 3 =
17
32 * 1 - (2 - 16) =
46
29 - 0 =
29
26 + 52 =
78
1 * 65 + (1 * 20) =
85
29 - 9 =
20
10 + 20 =
30
130 - 19 =
111
11 + (1 + 19) - 15 =
16
(52 / 13) + (68 + 23) =
95
65 - 18 =
47
22 - 0 =
22
4 * 18 + 16 - 24 =
64
66 - 10 =
56
22 + 4 + 10 - 1 =
35
0 / (29 * 17) =
0
18 + 35 * 19 =
683
27 + 12 + 50 =
89
77 + 5 =
82
30 + (13 - 19) + 18 =
42
28 + 23 =
51
47 + 8 =
55
(29 - 26) + 65 =
68
66 - 21 =
45
5 * 18 / 2 =
45
37 - 25 =
12
17 + 6 =
23
4 * (47 * 3) =
564
(14 * 144) / 28 - 64 =
8
(7 - 2) + 0 / 26 =
5
53 + 29 - 13 =
69
12 + 10 =
22
8 + 34 + 79 =
121
6 * (16 + 22) =
228
4 * 13 =
52
10 + 29 =
39
15 * 13 =
195
30 + 57 =
87
15 - 14 =
1
(59 * 6) - (6 * 1) =
348
99 / 11 =
9
(1 + 28) + (40 + 6) =
75
50 + 21 - 14 / 7 =
69
11 + (6 + 74) =
91
17 / 17 + 23 =
24
19 + 39 =
58
30 + 3 =
33
10 + 15 + 33 =
58
3 * 19 =
57
17 * 16 + 75 =
347
25 + 3 =
28
24 + 26 * 22 =
596
(88 / 2) - 23 =
21
0 * (27 + 47) + 46 =
46
26 / 26 + 11 + 11 =
23
(10 / 5) - 4 + 64 =
62
1 + 17 + 25 =
43
6 - 24 - 0 + 19 =
1
31 + 3 + 1 =
35
0 / 30 + 26 * 17 =
442
21 + (80 + 0) =
101
17 * 16 - 34 + 23 =
261
28 + 27 - 6 - 6 =
43
21 - 18 - 2 =
1
63 * 11 - 52 =
641
(0 - 4) + 20 * 21 =
416
End of preview. Expand in Data Studio

arithmark-2.0-train

The training set behind arithmark-2.0-scratch-19m, a small character level model trained from scratch that scores 97.40% on ArithMark 2.0.

It holds 3 million procedurally generated integer arithmetic problems in the same format as ArithMark 2.0. Each expression ends in "=" and has one correct answer. Division is always exact and answers are non negative.

Every one of the 2500 expressions in the ArithMark 2.0 benchmark was removed from this set, so a model trained here is still evaluated on problems it has not seen.

Format

{"ctx": "3 * 9 + 12 / 1 =", "answer": "39", "text": "3 * 9 + 12 / 1 = 39"}
  • ctx: the expression, ending in "="
  • answer: the correct integer answer as a string
  • text: ctx and answer joined, ready for language model training

Distribution

Problems follow the ArithMark 2.0 mix. About half are single operator, the rest use two or three operators, and roughly a quarter use parentheses. Operators are addition, subtraction, multiplication and exact division. Operand ranges match the benchmark.

License

Apache 2.0.

Downloads last month
56