Search papers, labs, and topics across Lattice.
University of Maryland, College Park, B-it, Probe-based steering requires roughly
5
0
7
1
Early hidden states of LLMs can predict steering success with surprising accuracy, enabling efficient steering without exhaustive rollouts.
Top systems in the ESDD2 challenge achieved a staggering Macro-F1 score of 0.8775, revealing the power of modular design and self-supervised learning in audio deepfake detection.
Dynamic order evolution in robotic warehousing can be tackled effectively, reducing order flowtime by leveraging cooperative strategies among robots.
Achieving up to 29.4% improvement in speech recognition accuracy under challenging conditions, M2S-AVSR redefines robustness in audio-visual speech tasks.
LLMs can exhibit surprising "strategic realism" when analyzing an ongoing geopolitical conflict, but their reasoning falters in politically ambiguous situations, revealing critical domain-specific limitations.
