元の投稿を読み込み中…X で見る ↗
Scam Signal Lens, open-source zero-shot scam-email classification eval tooling with replayable results, hit 100% precision/92% recall on English data but only ~17–23% recall on Spanish phishing, where localized questions and lower thresholds helped marginally.
元の投稿
元の投稿を読み込み中…X で見る ↗
jev
キーワードを入力してユースケースを検索。