1
Show HN: Zero-token memory retrieval for Pi
零token记忆检索方案，让Pi会话上下文精准回溯，兼顾图与时序视角。
Article URL: https://github.com/skorotkiewicz/zero-mem Comments URL: https://news.ycombinator.com/item?id=49218337 Points: 1 # Comments: 0
2
Empty Shelves or Lost Keys? Recall Is the Bottleneck for Parametric Factuality
揭秘大模型“记忆检索”瓶颈：参数化事实性缺失，究竟是模型没存住还是没找到？
arXiv:2602.14080v2 Announce Type: replace-cross Abstract: Standard factuality evaluations of LLMs treat all errors alike, obscuring whether failures a…
3
Towards Root Memories: Benchmarking and Enhancing Implicit Logical Memory Retrieval for Personalized LLMs
个性化大模型如何“懂你”？新研究为隐式逻辑记忆检索建立基准并提出增强方案，推动AI深层记忆能力进化。
arXiv:2606.23283v1 Announce Type: new Abstract: Memory systems are essential for personalized Large Language Models (LLMs). However, existing retrieva…
4
ActMem: Bridging the Gap Between Memory Retrieval and Reasoning in LLM Agents
新方法ActMem打通LLM Agent记忆检索与推理的断层，让智能体更聪明。
arXiv:2603.00026v2 Announce Type: replace Abstract: Memory management is essential for LLM agents in long-term interactions. Current memory frameworks…
5
Thinking Ahead: Prospection-Guided Retrieval of Memory with Language Models
提出前瞻性引导的记忆检索方法，让对话助手在长周期交互中准确找回语义相似度低的关键事实，突破传统RAG回溯局限。
arXiv:2605.14177v1 Announce Type: cross Abstract: Long-horizon personalization requires dialogue assistants to retrieve user-specific facts from exten…