$ open /benchmarks/scicode▌discussed░░░░░░░░░░░░░░░░░░░░ 0/65 problems · discussed · posts
SciCode
Science / Code
SciCode evaluates the ability to solve scientific research problems through code. Problems are decomposed into subproblems; this dev import preserves the relationships between 15 parent problems and 50 subproblems.
SciCode structures scientific coding problems as parent tasks with linked subproblems. Records provide descriptions, scientific background, input and output information, dependencies and, when supplied, code templates. The catalogue development selection contains 15 parent problems and 50 subproblems, giving 65 task records rather than 65 independent research problems. Reference solutions and test fields are kept outside the displayed prompts.
At a glance
Size
Catalog dev selection: 15 parent problems and 50 subproblems, totaling 65 task records.
License
apache-2.0 (declared on the Hugging Face dataset card)
Released
Dataset published on Hugging Face 2025-02
Maintainer
SciCode1 (Hugging Face organisation)
Answer access
Parent general_solution and child step_solution, ground_truth_code, and test_cases fields, when supplied, are kept outside the prompt in restricted records.
What a problem looks like
Parent scientific problems contain descriptions, background, input/output information, and dependencies. sub_steps become linked child tasks with step descriptions, background, input/output information, dependencies, and code templates when supplied.
For the 65 imported problems, the source ships an answer with every problem; this catalogue keeps it in restricted storage and never renders it.
Why it is hard
Requires solving scientific research problems through code, with parent problems decomposed into subproblems.
Limitations and criticism
Solving a supplied scientific coding task does not establish the ability to choose a research question or validate a scientific model against experimental evidence. Performance on an individual subproblem also does not show that a system can combine all the steps into a working parent solution. Results on this development selection should be distinguished from held-out evaluation, and parent and child outcomes should not be treated as independent observations.