Search papers, labs, and topics across Lattice.
1
0
3
A fine-tuned RoBERTa model matches the CVE-to-CWE classification accuracy of a model 64x larger, proving that smaller, specialized models can rival LLMs in niche tasks.
