Complete artifact for the empirical study "Memorised, Not Generated: Verbatim Recall of Published Fixtures in LLM-Generated Test Data, Measured Across Five Models and Eight Public Schemas" (submitted to Empirical Software Engineering). Version 1.2.1 adds experiment E1c: the copy-rate measurement (exact-text match, near...
Vijay Prasad Javvadi· Zenodo (CERN European Organi...· 0 citations
Reproducibility bundle for the manuscript "Trust, Augment, or Replace? What a Cross-Project Defect Predictor Is Worth on 1,267 Real Defects from Defects4J and BugsInPy" (Vijay Prasad Javvadi, submitted to PeerJ Computer Science, AI Application article type). The study asks whether a file-level defect predictor trained...
Vijay Prasad Javvadi· Zenodo (CERN European Organi...· 0 citations
A runtime for agentic software testing in which a language model proposes tool calls but decides nothing. Every proposed call is classified against a manifest and resolved by an operator-owned policy table into execute, ask, or refuse; an unclassified operation is refused by default. Completion is decided by code that...
Vijay Prasad Javvadi· Zenodo (CERN European Organi...· 0 citations
A runtime for agentic software testing in which a language model proposes tool calls but decides nothing. Every proposed call is classified against a manifest and resolved by an operator-owned policy table into execute, ask, or refuse; an unclassified operation is refused by default. Completion is decided by code that...
Vijay Prasad Javvadi· Zenodo (CERN European Organi...· 0 citations
A runtime for agentic software testing in which a language model proposes tool calls but decides nothing. Every proposed call is classified against a manifest and resolved by an operator-owned policy table into execute, ask, or refuse; an unclassified operation is refused by default. Completion is decided by code that...
Vijay Prasad Javvadi· Zenodo (CERN European Organi...· 0 citations
A runtime for agentic software testing in which a language model proposes tool calls but decides nothing. Every proposed call is classified against a manifest and resolved by an operator-owned policy table into execute, ask, or refuse; an unclassified operation is refused by default. Completion is decided by code that...
Vijay Prasad Javvadi· Zenodo (CERN European Organi...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
