Search papers, labs, and topics across Lattice.
1
0
3
2
LLM-generated authorial impersonations, despite their sophistication, are surprisingly detectable by existing authorship verification methods, even outperforming on some genuine negative samples.
