ホーム /ベンチマーク・評価・較正 /jev-playground by hegargarcia
jev-playground by hegargarcia
状態、合法な行動、測定可能な結果が明示されたゲームで、Jev と他のモデルを比較します。
このページは英語版リストから生成しています。
readme より
Jev Playground A playground for benchmarking TypeSafe AI’s Jev against other evaluation models in games with explicit states, legal actions, and measurable outcomes. The core question: how well does each model choose the next action when the rules and available choices are clearly defined? Games provide a small, inspectable environment for exploring …
詳細
- セクション
- ベンチマーク・評価・較正
- owner
- hegargarcia
- スター
- 0
- フォーク
- 0
- 最終更新
- 2026-09-17
- 言語
- TypeScript
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