ホーム /ベンチマーク・評価・較正 /jev-spam-eval
jev-spam-eval
Noul の質問によるゼロショットのスパムフィルタリングを TF-IDF のベースラインと比較します。
このページは英語版リストから生成しています。
readme より
Zero-shot email classification with TypeSafe’s Jev TypeSafe’s Jev reached 98.64% accuracy on a 5,733-email ham/spam/phishing test using written category definitions and email context, without task-specific fine-tuning or labeled examples in its requests. A TF-IDF logistic regression classifier trained on roughly 4,600 labeled messages per fold reached …
詳細
- セクション
- ベンチマーク・評価・較正
- owner
- bitnovus
- スター
- 0
- フォーク
- 0
- 最終更新
- 2026-09-18
- 言語
- Jupyter Notebook
- ライセンス
- MIT
これは何か
- 形態
- データセットやベンチマーク
- ホストエージェント
- 単体で動作
- 対象
- 研究者
成熟度
ドキュメント
実測値を報告しています