290,000 code snippets labeled with programming languages using automated text processing and language-specific patterns. The dataset excludes labels from ChatGPT or other LLMs, relying on pandas-based extraction and deterministic heuristics.
Use Cases
- Train a programming language identification model using the code snippet text and the associated language labels
- Benchmark the accuracy of LLM-based classifiers against these rule-based labels for language detection
- Analyze syntax patterns across different programming languages by filtering the dataset by the language label feature
Strengths
- 290,000 labeled code snippets sourced from Hugging Face
- Labels derived from language-specific patterns and automated text processing rather than LLMs
- Data prepared and manipulated using the pandas library for structured access