Search papers, labs, and topics across Lattice.
Affiliation:
3
0
7
The latest AI City Challenge reveals that integrating foundation models with synthetic data design can significantly boost performance in complex urban scenarios.
Hierarchical local attention in TextNCA reveals that the arrangement of attention windows can dramatically influence language modeling performance, even more than the model's iterative nature.
A comprehensive taxonomy reveals critical failure modes in LLM reasoning, exposing vulnerabilities that could hinder their deployment in real-world applications.
