Search papers, labs, and topics across Lattice.
3
0
6
3
Stop overpaying for LLM serving: intelligently routing requests to specialized pools based on token budget slashes GPU costs by up to 42% and dramatically improves reliability.
Even simple screen-level manipulations can trick computer-using agents into performing privileged actions, but a dual-channel guardrail offers a promising defense.
Stop blindly trusting LLM-generated code: STELP offers a secure execution environment that significantly outperforms existing methods in safely handling malicious code snippets.
