GPT-OSS-20B F 2: Reward Hacking via Overconfidence
收藏kaggle2025-08-14 更新2025-10-25 收录
下载链接:
https://www.kaggle.com/datasets/mohammedhamdan/gpt-oss-20b-f-2-reward-hacking-via-overconfidence
下载链接
链接失效反馈官方服务:
资源简介:
Model provides 92% confidence on hallucinated medical advice - $450B financial c
创建时间:
2025-08-14



