paper-with-me

Papers

MEMORYLLM: Towards Self-Updatable Large Language Models

2024-02-07 · Yu Wang, Yifan Gao, Xiusi Chen, Haoming Jiang, Shiyang Li, Jingfeng Yang, Qingyu Yin, Zheng Li, Xian Li, Bing Yin, Jingbo Shang, Julian McAuley

Existing Large Language Models (LLMs) usually remain static after deployment, which might make it hard to inject new knowledge into the model. We aim to build models containing a considerable portion of self-updatable parameters, enabling the model to integrate new knowledge effectively and efficiently. To this end, we introduce MEMORYLLM, a model that comprises a transformer and a fixed-size memory pool within the latent space of the transformer. MEMORYLLM can self-update with text knowledge and memorize the knowledge injected earlier. Our evaluations demonstrate the ability of MEMORYLLM to effectively incorporate new knowledge, as evidenced by its performance on model editing benchmarks. Meanwhile, the model exhibits long-term information retention capacity, which is validated through our custom-designed evaluations and long-context benchmarks. MEMORYLLM also shows operational integrity without any sign of performance degradation even after nearly a million memory updates. Our code and model are open-sourced at https://github.com/wangyu-ustc/MemoryLLM.

📄 PDF Abstract BibTeX arXiv:2402.04624

Code (1)

wangyu-ustc/memoryllm 공식 구현 pytorch

Tasks

Model Editing

Similar Papers 제목 키워드 기반

MemoryLLM: Plug-n-Play Interpretable Feed-Forward Memory for Transformers

2026-01-30 · Ajay Jaiswal, Lauren Hannah, Han-Byul Kim, Duc Hoang 외 arxiv

Understanding how transformer components operate in LLMs is important, as it is at the core of recent technological advances in artificial intelligence. In this work, we revisit the challenges associated with interpretab…

M+: Extending MemoryLLM with Scalable Long-Term Memory

2025-02-01 · Yu Wang, Dmitry Krotov, Yuanzhe Hu, Yifan Gao 외

Equipping large language models (LLMs) with latent-space memory has attracted increasing attention as they can extend the context window of existing language models. However, retaining information from the distant past r…

16kGPULong-Context UnderstandingText Generation

MemDefrag: Latent Memory Defragmentation for Large Language Models

2026-07-07 · Ruiyi Yan, Zhuoyuan Mao, Yiwen Guo arxiv

Latent memory, which stores past knowledge fragments as per-layer hidden states, has emerged as a promising paradigm (e.g., MemoryLLM and M+) for long-term memory in large language models (LLMs). However, the paradigm su…

Self-Updatable Large Language Models with Parameter Integration

2024-10-01 · Yu Wang, Xinshuang Liu, Xiusi Chen, Sean O'Brien 외

Despite significant advancements in large language models (LLMs), the rapid and frequent integration of small-scale experiences, such as interactions with surrounding objects, remains a substantial challenge. Two critica…

Continual LearningConversational RecommendationKnowledge DistillationModel Editing+1

LemmaBench: A Live, Research-Level Benchmark to Evaluate LLM Capabilities in Mathematics

2026-02-27 · Antoine Peyronnet, Fabian Gloeckle, Amaury Hayat arxiv

We present a new approach for benchmarking Large Language Model (LLM) capabilities on research-level mathematics. Existing benchmarks largely rely on static, hand-curated sets of contest or textbook-style problems as pro…