Search papers, labs, and topics across Lattice.
1
0
3
Current Large Audio Language Models (LALMs) struggle with basic audio understanding tasks like noise localization and cross-lingual speech, with some performing worse than random chance, despite excelling at speech recognition.
