#对齐技术

2 stories · 0 in 7 days · 1 in 30 days

AI助手 1 LLM安全 1 OpenAI 1 TRACE 1 多轮对话 1 研究实习生 1

Stories, products and papers tagged "对齐技术", newest first.

Oct 2

Papers Builders10:47

Chinese contentTRACE方法提升多轮对话安全性

TRACE让大模型在多轮对话中也能保持安全，性能损失极小，代码已开源。

#TRACE #LLM安全 #多轮对话 #对齐技术

arXiv cs.AI@Fengpeng Li et al. (6)Source

Sep 7

Models US vs ChinaMulti-sourceSelected73°21:28

OpenAI AI research interns

OpenAI's AI agents now complete over 3 days of work per human day, raising questions about AI safety and control as capabilities advance faster than safety measures.

#OpenAI #AI agents #research automation #AI safety

Event

官方账号Decoder@Maximilian Schreiner11 sourcesSource