Search papers, labs, and topics across Lattice.
3
0
7
2
CoReLoop enables additional refinement in an already-trained SSL-based detector while preserving the detector's original first-pass prediction by training only lightweight refinement modules and loop-specific low-rank adapters on the original data.
Loop Memory Attention enables models to revisit earlier computations, leading to a 2.2% accuracy boost in MathQA tasks compared to fixed loop depths.
SLMs that seem safe with text inputs can completely fail when the same content is spoken, revealing a critical "speech grounding gap" in current models.
