Search papers, labs, and topics across Lattice.
1
0
3
36
Visual context barely affects human sentence acceptability judgments, but throws LLMs for a loop, widening the gap between their internal representations and acceptability predictions.
