Jialun Cao

A study highlights data contamination's impact on code model evaluations.

A new benchmark to assess LLMs for Java programming tasks.

A new approach enhances testing reliability for deep learning libraries.

A multi-domain benchmark assesses LLMs' code generation abilities across various fields.

Introducing ADIT: A new approach to enhance software testing efficiency through automated input transformation.

Learn how code refactoring reduces data contamination in software development.