Search papers, labs, and topics across Lattice.
Iowa State University Ames
4
0
8
Adaptive sampling can significantly enhance LLM reasoning efficiency by tailoring resource allocation based on prompt difficulty and model confidence.
CudaPerf achieves up to 5X speedup in CUDA kernel generation by leveraging structural code-aware rewards that traditional methods overlook.
ReM-MoA reveals that structured cross-layer reasoning memory can dramatically enhance the scalability of multi-agent systems, outperforming traditional methods as complexity increases.
POTracker achieves a 51% boost in accuracy for generating compliant power outage reports, setting a new standard for domain-specific LLM applications.
