NAQD Env: A benchmark for selective withdrawal in language agents
·
1 次浏览
·来源: ArXiv AI
arXiv:2609.38460v1 Announce Type: new Abstract: Language agents must revise planned actions when evidence changes, permission is revoked, or a stop instruction arrives. A useful response is selective: suspend affected actions, preserve unaffected work, and resume only after sufficient repair. We int...
arXiv:2609.38460v1 Announce Type: new Abstract: Language agents must revise planned actions when evidence changes, permission is revoked, or a stop instruction arrives. A useful response is selective: suspend affected actions, preserve unaffected work, and resume only after sufficient repair. We int...
arXiv:2609.38460v1 Announce Type: new Abstract: Language agents must revise planned actions when evidence changes, permission is revoked, or a stop instruction arrives. A useful response is selective: suspend affected actions, preserve unaffected work, and resume only after sufficient repair. We int...