Search papers, labs, and topics across Lattice.
ISP RAS Research Center for Trusted Artificial Intelligence
2
0
5
Medical LLMs do not need to trade diagnostic rigor for bedside manner: conditioning generation on inferred patient affect consistently beats neutral baselines on both human clarity ratings and objective clinical benchmarks.
Deferring to a larger LLM only when a smaller LLM is uncertain can match the performance of the larger model alone, while slashing inference costs.
