Search papers, labs, and topics across Lattice.
Northeastern University
5
0
10
S-JEPA sets a new standard in speech representation learning by achieving top performance with fewer parameters and without the cumbersome offline re-clustering process.
StateGen achieves a remarkable 9.66/10 score in tool-call hallucination reduction, setting a new standard for synthetic data generation in multi-agent LLM training.
LLMs can slash token usage by 80% and "thinking rate" by 95% without sacrificing accuracy, simply by learning when *not* to reason.
Ditch the deterministic databases: this LLM-driven simulation framework evaluates tool-calling agents with surprisingly reliable proxy states, offering a scalable alternative to costly benchmarks.
LLM agents blindly trust third-party tools, but MCPShield adds a "security cognition" layer that learns to spot malicious servers before they can do damage.
