Search papers, labs, and topics across Lattice.
1
0
2
Aligning LLMs with human moral prototypes boosts their adversarial robustness while revealing critical flaws in existing alignment strategies.
