메이저리그 100년정도 데이터가지고 로스터만으로 예상 승률이랑 실제 승률 계산했더니 감독이 미치는 영향은 +-2승정도라고 함 그만큼 감독이 미치는 영향이 적단거임. 카드 얼마나 잘 구성하냐 싸움에 가깝지 전술적으로 뭔가 할 수 있는 여지가 작음
aai 벤치 최신버전에 오픈웨이트들 전부 점수 꼬라박아서 이럴거 같긴했는데
Kimi K3.1 recalls instructions it was never given on Terminal Bench. They match the hidden tests. K3.1 while solving: "Instructions likely distilled pitfalls from previous agents." 🤯
When we edit a task to contradict them, K3.1 refuses the new instructions and fails. Other models follow the edit.
Is it terminal(ly) benchmaxxed? 🧵
첨부엉 retweeted
Kimi K3.1 recalls instructions it was never given on Terminal Bench. They match the hidden tests. K3.1 while solving: "Instructions likely distilled pitfalls from previous agents." 🤯
When we edit a task to contradict them, K3.1 refuses the new instructions and fails. Other models follow the edit.
Is it terminal(ly) benchmaxxed? 🧵




