今日推介(第2205期)：大语言模型双用途知识的 Token 接种、利用特权信息
今日推介(第2205期)：大语言模型双用途知识的 Token 接种、利用特权信息
爱生活爱珂珂的文章
![[LG]《Exposure-Based Reinforcement Learni](http://dayu-img.uc.cn/columbus/img/oc/1002/7a8d26fa2fed1cc3313c9169542e9e3d.jpg)
[LG]《Exposure-Based Reinforcement Learni
[LG]《Exposure-Based Reinforcement Learni
![[CV]《Masked Visual Actions for Unified W](http://dayu-img.uc.cn/columbus/img/oc/1002/abe5e6b37b4e0f9fa95778a5b84ca9a6.jpg)
[CV]《Masked Visual Actions for Unified W
[CV]《Masked Visual Actions for Unified W
![[CL]《Search-on-Graph-R1: Training Large](http://dayu-img.uc.cn/columbus/img/oc/1002/7356b77a9789dfcc7a582773744b68b8.jpg)
[CL]《Search-on-Graph-R1: Training Large
[CL]《Search-on-Graph-R1: Training Large
![[LG]《Off-Context GRPO: Learning to Reaso](http://dayu-img.uc.cn/columbus/img/oc/1002/23a52b0ad7c48e7487fcfb1748027db6.jpg)
[LG]《Off-Context GRPO: Learning to Reaso
[LG]《Off-Context GRPO: Learning to Reaso
![[LG]《Mark, Don't Erase: Token Inoculatio](http://dayu-img.uc.cn/columbus/img/oc/1002/71cf9d1d4d33ba05bea12b10737a1ee2.jpg)
[LG]《Mark, Don't Erase: Token Inoculatio
[LG]《Mark, Don't Erase: Token Inoculatio
早！ 早安
早！ 早安
晚安～ 晚安
晚安～ 晚安
一套面向生产级的 Pi-Agent SDK 深度学习笔记，系统拆解了 10 章源
一套面向生产级的 Pi-Agent SDK 深度学习笔记，系统拆解了 10 章源
人工智能 文科生 这次不是文科的胜利，是“会跟AI说人话”的人的胜利。那些以为自
人工智能 文科生 这次不是文科的胜利，是“会跟AI说人话”的人的胜利。那些以为自
![[LG]《Can We Break LLMs Out of Self-Loops](http://dayu-img.uc.cn/columbus/img/oc/1002/ce4422c641e8f37b2e23c720a032d04a.jpg)
[LG]《Can We Break LLMs Out of Self-Loops
[LG]《Can We Break LLMs Out of Self-Loops
![[LG]《Discovery by Dreaming: Cross-Domain](http://dayu-img.uc.cn/columbus/img/oc/1002/adb5c23633ea436170ba9a28d7bf2389.jpg)
[LG]《Discovery by Dreaming: Cross-Domain
[LG]《Discovery by Dreaming: Cross-Domain
![[AI]《Fantastic Adaptive Taxonomies and H](http://dayu-img.uc.cn/columbus/img/oc/1002/ace8fa96e3f555aac20a8c7a5023c4a3.jpg)
[AI]《Fantastic Adaptive Taxonomies and H
[AI]《Fantastic Adaptive Taxonomies and H
![[RO]《Patch Policy: Efficient Embodied Co](http://dayu-img.uc.cn/columbus/img/oc/1002/e8d92c9fc78e51ecdc1f82bfa6623b9a.jpg)
[RO]《Patch Policy: Efficient Embodied Co
[RO]《Patch Policy: Efficient Embodied Co
![[LG]《LLM-as-a-Coach: Experiential Learni](http://dayu-img.uc.cn/columbus/img/oc/1002/fa6a5f9f07cacb671927bf01d9d9d0bd.jpg)
[LG]《LLM-as-a-Coach: Experiential Learni
[LG]《LLM-as-a-Coach: Experiential Learni
早！ 早安
早！ 早安
晚安～ 晚安
晚安～ 晚安
【如果你想靠自学拿下一个数学学位：一份清醒的通关指南】Susan Rigetti
【如果你想靠自学拿下一个数学学位：一份清醒的通关指南】Susan Rigetti
《深入理解 AI Agent：设计原理与工程实践》已开源全书正文、编译版 PDF
《深入理解 AI Agent：设计原理与工程实践》已开源全书正文、编译版 PDF
AI的六大美德 这里有个悖论：你越享受AI“接住你所有情绪”，你在现实中能被接住
AI的六大美德 这里有个悖论：你越享受AI“接住你所有情绪”，你在现实中能被接住
作者信息
