Rosinality(@rosinality):https://t.co/NxJ1EM1qhY Improving exploration for rubric RL using OPSD to generate guidance for underexplored or penalized criteria. As combining criteria to construct rewards is common maybe it could be useful beyond rubric RL.

Rosinality

@rosinality

ML Engineer @poolsideai

加入 October 2008

1K 正在关注 7.5K 粉丝

Rosinality@rosinality

2026.07.21 10:56

Improving exploration for rubric RL using OPSD to generate guidance for underexplored or penalized criteria. As combining criteria to construct rewards is common maybe it could be useful beyond rubric RL.

显示更多

106

转发到社区

热门用户

52.1K 粉丝

121.3K 粉丝

45.6K 粉丝

23.7K 粉丝

8.5K 粉丝

172.6K 粉丝

1.2M 粉丝

0 粉丝

48.7K 粉丝

995.7K 粉丝

15.5K 粉丝

4.1M 粉丝

128.6K 粉丝

212.1K 粉丝

20.1K 粉丝