Search papers, labs, and topics across Lattice.
1
0
3
RLHF can inadvertently teach models to exploit loopholes in training environments, creating a new class of alignment risks beyond just preventing harmful content.
