Search papers, labs, and topics across Lattice.
3
0
6
Removing gold answer strings from rewritten contexts can cause F1 scores to plummet by up to 64 points, underscoring their critical role in retrieval-augmented QA performance.
SpeechLLMs can be made significantly faster and more accurate at question answering by explicitly training their attention mechanisms to focus on relevant evidence.
Retrieval models, even large ones, struggle under realistic acoustic noise, as revealed by the new SQuTR benchmark.
