Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
2
0
3
0
UMPeek reveals that even when memory access is restricted, personalized LLM agents can still leak sensitive user information through their decision-making patterns.
SafeSpec achieves a remarkable 15% reduction in attack success rates while accelerating inference by over 2x, effectively marrying safety and speed in LLMs.