Diagnostic Foundation for Evaluating LLMs' Research Integrity as Co-Scientists

arXiv:2608.12345v1 Announce Type: new Abstract: Language models are increasingly deployed as coscientists, yet their ability to uphold research integrity under institutional pressure remains unmeasured.