Clinical Context Variables Collectively Rival Model Choice in Embedding-Based Retrieval: Multi-Corpus Benchmark Study
Yngve Mikkelsen, Arriel Benis, Katharina Danhauser, Yunguo Yu · 2026
Read Clinical Context Variables Collectively Rival Model Choice in Embedding-Based Retrieval: Multi-Corpus Benchmark Study by Yngve Mikkelsen, Arriel Benis, Katharina Danhauser, Yunguo Yu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13195371.txt.
Description
Yngve Mikkelsen and co-authors benchmark clinical document search systems in a 2026 study. Search quality decides whether AI assistants find the right evidence at all. Thirteen search setups ran across three clinical text sets and two query styles. Four chunking plans and many scoring metrics filled out the design.
Model choice explained about forty-one percent of score spread. Corpus type and query format together explained a like share, near half with interactions. Chunk size barely mattered at all. Rankings wobbled under keyword queries but held firm under natural ones. Domain tuned models oddly trailed general ones here. Setup context rivals model choice, the authors conclude.
Catalog record
| CREATOR | Yngve Mikkelsen, Arriel Benis, Katharina Danhauser, Yunguo Yu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 74 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |