Dan Stillman a5fd13e2a4 Subtract each model's mean vector before comparing embeddings ... Every embedding a model produces shares a large common direction that says nothing about the text, so items with little content scored as moderate matches against everything: for the query "sun", an item titled "C" outscored a paper about a coronal mass ejection. Removing that direction spreads the scores out, so no relevance reads as no score. The mean is a constant per model, computed over titles and abstracts across fields and languages, and applies to stored vectors as they're compared, so the index doesn't change. Display ranges are refit to the scores that result.		2026-08-05 15:41:42 -04:00
..
components	fx140: Asyncify/ESMify tests	2025-07-30 22:30:53 -04:00
content	fx153: ownerGlobal -> documentGlobal	2026-08-03 11:48:40 -04:00
resource	Remove Chai as Promised	2025-07-30 22:31:08 -04:00
tests	Subtract each model's mean vector before comparing embeddings	2026-08-05 15:41:42 -04:00
chrome.manifest	fx115: Restore test runner	2024-03-30 00:58:53 -04:00
runtests.sh	runtests.sh: Add -p option to run a shard of the test files (e.g., -p 2/4)	2026-07-08 10:50:32 -04:00