正在加载视频...

视频加载失败

🔭 We’re releasing Hodoscope: an open-source tool for unsupervised behavior discovery. It lets you visually explore and compare agent behaviors at scale. It helped us discover a novel reward hacking vulnerability in Commit0 - with just a couple minutes of human effort.

74,566 次观看 • 3 个月前 •via X (Twitter)

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频