Search papers, labs, and topics across Lattice.
2
0
4
Negation detection in multimodal systems is fundamentally flawed, with pretrained models missing a generalizable signal, but a new attention architecture bridges this gap, enhancing performance significantly.
LLMs can internally compute the right answer to simple reasoning tasks like character counting, but then actively suppress that information before outputting the final (wrong) answer.
