TypeSafe AIが引用チェッカーに追加
ユーザーが引用チェッカーに@typesafeaiを追加し高評価。PaperTrellisで無料試用可能。Jevと医療AIのタグ付き。 Link
Wasim Jalali 氏が Decision Bench 9日目の結果を共有:Clef Omni は87.2%の精度で8モデル中最下位。Jev 1.13 と隠れた質問数やノイズ処理を比較。保守的で大半のノイズを残す。
Decision Bench, day 9: Clef Omni lands at #8 of 8. • 87.2% accuracy across 1,734 checks, working alone • 3 real questions hidden, against 1 for Jev 1.13 • 219 noise comments left in, about $0.431 for the whole run It plays very safe and leaves most of the noise in. Hiding