Search papers, labs, and topics across Lattice.
Renmin University of China
2
0
5
Major LLM failures can be traced back to specific disruptions in cognitive organization, providing a roadmap for targeted interventions and deeper understanding of model behavior.
Current vision-language models struggle to identify unsafe actions in dynamic household environments, highlighting a critical gap that HomeSafe-Bench aims to fill.
