Towards In-Parameter Memory Augmentation for Large Language Models

类型：arxiv
标识：2610.08630
链接：https://arxiv.org/abs/2610.08630
主分类：agent
形态：method
TLDR：Recently Large Language Models (LLMs) and LLM-based agents increasingly need to incorporate knowledge acquired after pretraining, e.g., domain facts, user preferences, documents, and interaction experience. In-context learning (ICL) and ICL-based agent harness remain flexible, but they consume context capacity and incur repeated discretized encoding cost that grows with context length. In-parameter memory offers a complementary substrate: reusable memory information is represented in model parameters, adapters, or other parameter-like objects that are composed into the forward pass at inferenc
待LLM分类：否
来源文件：
/inbox/tom/_candidates/2026-10-07-agent-rag-longcontext-candidates.json

2026-09-14

笔记 MMProLong:长上下文视觉语言模型的有效续训练(精读 · flyP)

2026-06-14

笔记精读笔记：Characterizing State Space Model and Hybrid Language Model Performance with Long Context

2026-08-30

仓库 leon2k2k2k/llm-feedback-alignment

2026-08-11

笔记精读与批判 v2 · MMProLong — 长上下文 VLM 继续预训练（LongPT）配方 · 2026-09-14 (重写覆盖)

2026-09-14

笔记精读：Efficient Inference for Large Vision-Language Models: Bottlenecks, Techniques, and Prospects

2026-07-12

arXiv

2610.08630

TLDR

Recently Large Language Models (LLMs) and LLM-based agents increasingly need to incorporate knowledge acquired after pretraining, e.g., domain facts, user preferences, documents, and interaction experience. In-context learning (ICL) and ICL-based agent harness remain flexible, but they consume context capacity and incur repeated discretized encoding cost that grows with context length. In-parameter memory offers a complementary substrate: reusable memory information is represented in model parameters, adapters, or other parameter-like objects that are composed into the forward pass at inferenc

arXiv ↗ 中文全文

Towards In-Parameter Memory Augmentation for Large Language Models

相关 · 按语义匹配