Large language models can recover author identities from titles and abstracts even when stylistic and bibliographic cues are removed. Models identify authors by detecting latent conceptual signatures in problem framing and research focus, effectively collapsing the anonymity of double-blind reviews.
HOW THIS AFFECTS YOU
●
researcherYour research focus and problem framing could inadvertently reveal your identity.
●
policyYou may need to rethink anonymization standards for scientific publishing.