Execugraph Internal30: 30 Data Structures and Algorithms Benchmark Problems

Anonymized for double-blind review, this dataset is a curated suite of 30 data-structures-and-algorithms problems. It serves as the headline benchmark for the ExecuGraph framework, with each problem including a natural-language specification and function signature. The dataset was last updated on 2026-06-03.

Use Cases

• Benchmarking algorithm synthesis systems based on the curated suite of 30 problems.
• Evaluating code generation models based on natural-language problem specifications.
• Training models for automated programming tasks based on unambiguous function signatures.
• Studying the performance of data-structures-and-algorithms problem solvers.

Strengths

• Contains 30 curated data-structures-and-algorithms problems.
• Each problem includes an unambiguous natural-language specification and target function signature.
• Serves as the headline benchmark for the ExecuGraph framework.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

huggingface

Collection Method

Curated suite of benchmark problems.

Freshness

Last updated 2026-06-03 18:45:48; freshness should be verified.