What it does
This is a single historical search test, not a medical discovery or treatment recommendation. Modern training and MeSH indexing leave leakage risks, and the larger hidden-years benchmark remains planned.
Benchmarks & research
Jev Discovery
A historical PubMed experiment classifies literature links and ranks candidate substances in code, explicitly reporting that it missed its fish-oil rediscovery target.
Bookmark: Jev Discovery
Keep this in your collection.
Leave a noteWhat would you try with this?
: Jev Discovery
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
How you can use it
Study the historical question first: could papers published before 1986 connect Raynaud's syndrome with fish oil? Jev identifies changes in body measures in disease papers, then judges how substances affect those measures in other papers. Code counts matching directions to rank possible links.
Compare that ranking with the stated target, rather than treating a high place as a medical finding. The reported fish-oil candidate finished 74th, missing the top-ten goal. Modern model training and paper indexing can leak later knowledge; this search study is not treatment advice.