Search papers, labs, and topics across Lattice.
Beijing University of Posts and Telecommunications
2
0
5
LLM-based multi-agent systems are surprisingly vulnerable: a new RL-based attacker can evolve sophisticated, long-horizon attacks by exploiting trust in external tools.
Achieve SOTA reasoning performance with a 30B parameter model using up to 95% fewer tokens by explicitly controlling when and how deeply an LLM plans.
