IssueSpecter combines coverage analysis with LLM defect detection to generate prioritized, actionable issue reports, achieving 84.6% validity on manually reviewed issues from 13 Python projects and outperforming a coverage-driven baseline.
Evaluating and improving chatgpt for unit test generation.Proc
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.SE 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
LLMs produce counterexamples that remove up to 11.68% of invalid assertions from dynamic inference and raise precision by up to 7% on benchmarks without hurting recall.
citing papers explorer
-
LLM-Guided Issue Generation from Uncovered Code Segments
IssueSpecter combines coverage analysis with LLM defect detection to generate prioritized, actionable issue reports, achieving 84.6% validity on manually reviewed issues from 13 Python projects and outperforming a coverage-driven baseline.
-
Improving Dynamic Specification Inference with LLM-Generated Counterexamples
LLMs produce counterexamples that remove up to 11.68% of invalid assertions from dynamic inference and raise precision by up to 7% on benchmarks without hurting recall.