Search papers, labs, and topics across Lattice.
2
0
5
4
Current document parsing methods fail to consistently extract semantically meaningful information from enterprise documents, hindering their utility for AI agents.
LLMs can deanonymize authors with surprising accuracy, but this work offers a framework to understand and mitigate these risks through stylometric analysis and text rewriting.
