一作ACL 2026Findings· Findings of ACL
Beyond Superficial Tests: Adversarial Refinement for Reliable Property-Based Testing
Xiao Li · Runlin Liu · Zhe Zhang · Xiang Gao · Hailong Sun
摘要
Large Language Models (LLMs) have demonstrated remarkable proficiency in code generation, yet their application to Property-Based Testing (PBT) remains fraught with a superficiality gap. While LLMs can readily generate syntactically correct tests, they often struggle to bridge the semantic gap be...