-
TEGCER Dataset
A dataset of 15,000+ erroneous student programs which fail to compile, along with their corresponding repaired code. -
MBPP: A Dataset for Language Model Evaluation
The MBPP dataset is a collection of basic programming questions used to evaluate the performance of language models. -
APPS: A Dataset for Code Generation Evaluation
The APPS dataset is a collection of programming problems used to evaluate the performance of code generation models.