影响力指数

-/100

发表论文2 篇

平均评分4.5

年均产出2.0 篇/年

Nevan Wichers

Fellow@Anthropic·美国·OpenReview

Inoculation Prompting: Instructing LLMs to misbehave at train-time improves test-time alignment

ICLR 2026Rejected

Re-contextualization Mitigates Specification Gaming without Modifying the Specification

ICLR 2026Rejected

合作者 (15)

Alexander Matt Turner

Bryce Woodworth

Vladimir Ivanov