https://cognaptus.com/tags/stale-preferences/ Recent content in Stale-Preferences on Cognaptus Hugo -- 0.145.0 en-us Sun, 27 Sep 2026 00:00:00 +0000 https://cognaptus.com/blog/2026-09-27-yesterdays-gold-todays-bias-reusing-human-feedback-after-a-model-upgrade/ Sun, 27 Sep 2026 00:00:00 +0000 https://cognaptus.com/blog/2026-09-27-yesterdays-gold-todays-bias-reusing-human-feedback-after-a-model-upgrade/ StalePO shows how production translation teams may recover useful corrections from legacy post-edits without training an upgraded model to imitate an older system.