Search papers, labs, and topics across Lattice.
University of Sydney
1
0
2
Linear transformers can achieve efficient in-context learning by mapping context distributions to response functions, challenging the traditional softmax approach.
