shhu2001 commited on
Commit
4535fe4
·
verified ·
1 Parent(s): 27d45d8

Add SciCode-Verified dataset card

Browse files
Files changed (1) hide show
  1. README.md +107 -0
README.md ADDED
@@ -0,0 +1,107 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ pretty_name: SciCode-Verified
3
+ license: apache-2.0
4
+ language:
5
+ - en
6
+ annotations_creators:
7
+ - expert-generated
8
+ source_datasets:
9
+ - extended
10
+ task_categories:
11
+ - text-generation
12
+ size_categories:
13
+ - n<1K
14
+ tags:
15
+ - code
16
+ - science
17
+ - benchmark
18
+ - code-generation
19
+ - scientific-coding
20
+ - llm-evaluation
21
+ configs:
22
+ - config_name: default
23
+ data_files:
24
+ - split: test
25
+ path: data/problems_test.jsonl
26
+ ---
27
+
28
+ # SciCode-Verified
29
+
30
+ SciCode-Verified is the corrected, human-verified release of the
31
+ [SciCode](https://github.com/scicode-bench/SciCode) scientific-code-generation benchmark.
32
+ A problem-by-problem audit identified 263 defects in the 65-problem SciCode test split and
33
+ corrected every confirmable defect. The released evaluation set contains 64 main problems and
34
+ 287 scored subproblems; one original problem is excluded because its specification does not
35
+ determine a unique, verifiable answer.
36
+
37
+ - **Paper:** [SciCode-Verified: How Benchmark Defects Underestimated the Scientific-Coding Ability of Language Models](https://arxiv.org/abs/2608.04975)
38
+ - **Code, evaluation harness, and audit trail:** [github.com/flyingwagner/scicode-verified](https://github.com/flyingwagner/scicode-verified)
39
+ - **Original benchmark:** [SciCode](https://arxiv.org/abs/2407.13168)
40
+
41
+ ## Files
42
+
43
+ | File | Purpose |
44
+ |---|---|
45
+ | `data/problems_test.jsonl` | Corrected prompts and problem specifications, one main problem per line |
46
+ | `test_data_cleaned.h5` | Corrected frozen grading targets used by the evaluation harness |
47
+ | `manifest.json` | Release version, problem order, and MD5 checksums |
48
+ | `LICENSE` | Apache License 2.0 inherited from the upstream benchmark |
49
+
50
+ Release `v2` checksums:
51
+
52
+ ```text
53
+ 5c604d8dbf52642bd94e13b92c8f52eb data/problems_test.jsonl
54
+ 2b41a7df40ddc23ce651ec05b8ecb6f8 test_data_cleaned.h5
55
+ ```
56
+
57
+ ## Load the problem specifications
58
+
59
+ ```python
60
+ from datasets import load_dataset
61
+
62
+ dataset = load_dataset("shhu2001/SciCode-Verified", split="test")
63
+ print(dataset[0])
64
+ ```
65
+
66
+ ## Download the grading targets
67
+
68
+ ```python
69
+ from huggingface_hub import hf_hub_download
70
+
71
+ h5_path = hf_hub_download(
72
+ repo_id="shhu2001/SciCode-Verified",
73
+ repo_type="dataset",
74
+ filename="test_data_cleaned.h5",
75
+ )
76
+ ```
77
+
78
+ The full generation, grading, multi-environment evaluation, and integrity-check workflow is
79
+ documented in the [GitHub repository](https://github.com/flyingwagner/scicode-verified). Model
80
+ outputs and evaluation logs are intentionally not included in this dataset repository.
81
+
82
+ ## License and attribution
83
+
84
+ SciCode-Verified is derived from SciCode and redistributed under the Apache License 2.0. Please
85
+ cite both the original SciCode benchmark and SciCode-Verified.
86
+
87
+ ```bibtex
88
+ @article{hu2026scicodeverified,
89
+ title = {{SciCode-Verified}: How Benchmark Defects Underestimated the Scientific-Coding Ability of Language Models},
90
+ author = {Hu, Sihan and Huang, Lyuhan and Deng, Youjin and Chen, Kun},
91
+ year = {2026},
92
+ eprint = {2608.04975},
93
+ archivePrefix = {arXiv},
94
+ primaryClass = {cs.SE},
95
+ url = {https://arxiv.org/abs/2608.04975}
96
+ }
97
+
98
+ @article{tian2024scicode,
99
+ title = {{SciCode}: A Research Coding Benchmark Curated by Scientists},
100
+ author = {Tian, Minyang and Gao, Luyu and Zhang, Shizhuo Dylan and Chen, Xinan and others},
101
+ year = {2024},
102
+ eprint = {2407.13168},
103
+ archivePrefix = {arXiv},
104
+ primaryClass = {cs.AI},
105
+ url = {https://arxiv.org/abs/2407.13168}
106
+ }
107
+ ```