Search papers, labs, and topics across Lattice.
Affiliation:
4
23
8
8
Sparse updates from surgical alignment can boost reasoning quality in LLMs, even when accuracy takes a hit.
Forget slow, uninterpretable KG lookups: LogosKG unlocks interactive, billion-edge knowledge graph retrieval with orders-of-magnitude speedups by compiling symbolic queries to hardware-optimized operations.
LLMs can provide reliable probabilities for binary classification *without* sacrificing their ability to generate high-quality, aligned explanations, thanks to a novel dual-head fine-tuning approach.
Forget slow, expensive human evaluations: LLMs can reliably judge the quality of AI-generated medical summaries, achieving near-human agreement in seconds.
