← 返回归档

帖子详情

查看 X 原文
O
elvis@omarsar0
// Latent Semantic Memory for LLMs // This paper presents some really interesting findings about agent memory. They find that sixteen memory tokens beat full-context inference on FactKG while using 108x fewer tokens for the source documents. Approach: Documents are encoded once into reusable embeddings. For each query, a learned module gathers the relevant information from the whole collection and writes it into 16 tokens in the input space of a frozen LLM. Results: It beats Mem0, MemoRAG, and other baselines on five of eight benchmarks. On FactKG, it scores 84.33% against 80.43% with the full context. It also works as agent memory. In TextWorld, using the agent's own history as the collection raises success from 18.6% to 24.2%, and the agent revisits fewer rooms. It also seems that it could help frozen LLMs accumulate and use knowledge effectively across long-context interactions. This is applicable to building continually learning AI systems. arxiv.org/abs/2610.21375 Chat with Paper: academy.dair.ai/papers/memoril…
媒体保存失败：[Errno 104] Connection reset by peer
帖子 ID
2111879742454665582
Markdown
/app/data/archive/2026/10/18/omarsar0/2111879742454665582/post.md
保存状态
complete
查看 Markdown 内容
---
tweet_id: "2111879742454665582"
author: "omarsar0"
published_at: "2026-10-18T17:59:02+00:00"
source_url: "https://x.com/omarsar0/status/2111879742454665582"
---

// Latent Semantic Memory for LLMs //  
  
This paper presents some really interesting findings about agent memory.  
  
They find that sixteen memory tokens beat full-context inference on FactKG while using 108x fewer tokens for the source documents.  
  
Approach:  
  
Documents are encoded once into reusable embeddings. For each query, a learned module gathers the relevant information from the whole collection and writes it into 16 tokens in the input space of a frozen LLM.  
  
Results:  
  
It beats Mem0, MemoRAG, and other baselines on five of eight benchmarks. On FactKG, it scores 84.33% against 80.43% with the full context.  
  
It also works as agent memory. In TextWorld, using the agent's own history as the collection raises success from 18.6% to 24.2%, and the agent revisits fewer rooms.  
  
It also seems that it could help frozen LLMs accumulate and use knowledge effectively across long-context interactions. This is applicable to building continually learning AI systems.  
  
[arxiv.org/abs/2610.21375](https://arxiv.org/abs/2610.21375)  
  
Chat with Paper: [academy.dair.ai/papers/memoril…](https://academy.dair.ai/papers/memorilla-latent-semantic-memory-for-llms-2610.21375)

![Tweet Image](https://pbs.twimg.com/media/HU7nbt2aEAAKl18.jpg)