Search papers, labs, and topics across Lattice.
McGill University, LawZero
1
0
2
Adversarial training on just 66 carefully chosen statements can achieve near-zero attack success rates, revolutionizing safety alignment in language models.
