Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
Department of Computer Science and Engineering, Ewha Womans University
3
0
6
Search-augmented LLMs frequently cite real sources in ways that mislead users, creating a fidelity-suitability trade-off where faithful models select inappropriate sources and vice versa.
Initializing prompts in flatter regions of the loss landscape dramatically improves calibration and performance in test-time prompt tuning for vision-language models.
VLMs learn faster and better when you dynamically weight the prefixes based on input token importance, rather than treating all tokens equally.