Search papers, labs, and topics across Lattice.
Affiliation:
2
0
5
22
Control experiments on three cybersecurity datasets covering MITRE ATT&CK network traffic, cyber threat intelligence (CTI), and CVE vulnerability analysis show that reasoning topology substantially affects performance, and Graph reasoning achieves the highest overall accuracy.
Unlock up to 40% reasoning improvements in smaller LLMs simply by structuring your prompts.
