Private eval: pplx-decider shows 40% recall gain over Jev decision model

marktenenholtz · x · 2026-10-20

Mark Tenenholtz reports that on a private eval set for a personal project, pplx-decider showed a 40% increase in recall over Jev, concluding that decision models "have a long way to go." No further details about the eval set were shared.

Related event: pplx-decider Beats Jev by 40% Recall in Unofficial Test, But Decision Models Still Immature(2 posts)→

Original post →

More from Models

Models channel →