Search papers, labs, and topics across Lattice.
University of Southern California
2
0
3
ORBIT achieves superior multi-attribute steering in language models without the need for retraining, overcoming the limitations of previous methods that struggle with norm imbalance and directional cancellation.
Naively applying gRPC to cross-silo federated learning with large models can be a bottleneck; gRPC+S3 offers a surprisingly effective solution.
