Seven LLMs detect documentation faults far more reliably than implementation drift; when only the code changes and Javadoc stays intact, inconsistency detection falls by 21–43 percentage points.
Docprism: Local categorization and external filtering to identify relevant code-documentation inconsistencies,
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.SE 2years
2026 2representative citing papers
An empirical study extracts 20,729 expected behaviors from ten Java libraries and finds 17.5% remain untested, independent of line coverage and mutation scores.
citing papers explorer
-
Measuring LLM Trust Allocation Across Conflicting Software Artifacts
Seven LLMs detect documentation faults far more reliably than implementation drift; when only the code changes and Javadoc stays intact, inconsistency detection falls by 21–43 percentage points.
-
Beyond Coverage and Kill Scores: Empirically Measuring Test Suite Behavioural Gaps
An empirical study extracts 20,729 expected behaviors from ten Java libraries and finds 17.5% remain untested, independent of line coverage and mutation scores.