Search papers, labs, and topics across Lattice.
1
0
2
5
Interleaved speech-text models can implicitly transcribe spoken words into text with up to 77% accuracy, revealing a hidden layer of understanding in multimodal processing.
