pipette
ENEnglish

Verification Turns Capability Gaps into Routing Decisions: Small Language Models for Clinical Documentation

Nina Fatehi, Meysam Ghaffari, Animesh Agarwal, Anushka Swarup, Carlos Morato

PreprintUso en el mundo real

En palabras de los autores

Clinical note generation from doctor--patient dialogue is high-volume and hallucination-intolerant: a fabricated laboratory value can create patient-safety risk, not a quality regression. We ask whether an agentic verification architecture, rather than generator scale, can make a small language model viable here. Holding a four-evaluator verification stack and refinement policy fixed and swapping only the generator between a 4B-parameter model with a QLoRA adapter and a frontier baseline, the small model reaches claim-level factual entailment against and comes within points on completeness, generating entirely on-device. Its deficit is confined to inter-section structural fidelity---arrangement rather than clinical comprehension, and the axis on which it also falls furthest below the corpus's reference notes. Because that deficit is an explicit verifier decision rather than a silent quality difference, it becomes a routing policy: run the small model by default, escalate what the verifier rejects. Measured end to end, that it achieved a similiar verified-rate acceptance to the frontier-only acceptance in this test set ( vs.\ over encounters) while routing , documenting three quarters without a frontier generation call. Ablating the loop separates architecture from scale: unrefined, the small model is accepted on of encounters and the frontier model on , so verification does the most work where the generator is weakest. We also contribute section-scoped refinement, which preserves unflagged sections by construction rather than instruction, and find the baseline is not uniformly stronger---its provider refused two trauma-related transcripts the on-device model handled and certified, one facet of a provider dependency a served adapter does not carry.

Resultado principalLimitación que admiten los autores

Apareció: jueves, 15 de octubre. arXiv. Preprint, todavía sin revisión por pares.