Search papers, labs, and topics across Lattice.
Corresponding author
1
0
EVIL-Detect outperforms existing methods by achieving a macro-F1 score of 0.8888 in detecting LLM-generated text, even under challenging conditions.
