Search papers, labs, and topics across Lattice.
The Hong Kong University of Science and Technology, Hong Kong, China, Carnegie Mellon University
3
0
6
10
Achieving trajectory-level differential privacy in adaptive streaming contexts without sacrificing performance is now feasible through an auditable buffering-aggregation approach.
MIRA achieves superior mid-training data selection by dynamically constructing source-specific evaluation rubrics, outperforming traditional methods while using half the data.
Shadow API audits reveal that some premium Claude endpoints are statistically inconsistent with their reference models, raising concerns about model misrepresentation in LLM APIs.
