홈 /벤치마크, 평가, 보정 /jev-playground by hegargarcia
jev-playground by hegargarcia
명시적 상태, 합법 행동, 측정 가능한 결과를 갖춘 게임에서 Jev를 다른 모델과 비교합니다.
이 페이지는 영문 목록에서 생성했습니다.
readme에서
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 …
상세
- 섹션
- 벤치마크, 평가, 보정
- owner
- hegargarcia
- 스타
- 0
- 포크
- 0
- 최근 커밋
- 2026-09-17
- 언어
- TypeScript
어떤 프로젝트인가
- 형태
- 데이터셋 또는 벤치마크
- 호스트 에이전트
- 단독 실행
- 대상
- 연구자
성숙도
문서