# δ-mem: Efficient Online Memory for Large Language Models

arXiv: https://arxiv.org/abs/2605.12357 · HTML: https://mlsanity.com/paper/2605.12357/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- Builds on 37 of its 44 references that are in the map.
- First posted on arXiv 12 May 2026.

## Builds on (most cited first, 30 of 37)

- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [Prefix-Tuning: Optimizing Continuous Prompts for Generation](https://mlsanity.com/paper/2101.00190/) (2021)
- [HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering](https://mlsanity.com/paper/1809.09600/) (2018)
- [Generative Agents: Interactive Simulacra of Human Behavior](https://mlsanity.com/paper/2304.03442/) (2023)
- [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://mlsanity.com/paper/2311.12022/) (2023)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [MemGPT: Towards LLMs as Operating Systems](https://mlsanity.com/paper/2310.08560/) (2023)
- [Improving language models by retrieving from trillions of tokens](https://mlsanity.com/paper/2112.04426/) (2021)
- [Instruction-Following Evaluation for Large Language Models](https://mlsanity.com/paper/2311.07911/) (2023)
- [Mass-Editing Memory in a Transformer](https://mlsanity.com/paper/2210.07229/) (2022)
- [Evaluating Very Long-Term Conversational Memory of LLM Agents](https://mlsanity.com/paper/2402.17753/) (2024)
- [Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory](https://mlsanity.com/paper/2504.19413/) (2025)
- [MemoryBank: Enhancing Large Language Models with Long-Term Memory](https://mlsanity.com/paper/2305.10250/) (2023)
- [A Dataset of Information-Seeking Questions and Answers Anchored in Research Papers](https://mlsanity.com/paper/2105.03011/) (2021)
- [LLMs Get Lost In Multi-Turn Conversation](https://mlsanity.com/paper/2505.06120/) (2025)
- [Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention](https://mlsanity.com/paper/2502.11089/) (2025)
- [LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compression](https://mlsanity.com/paper/2403.12968/) (2024)
- [Titans: Learning to Memorize at Test Time](https://mlsanity.com/paper/2501.00663/) (2024)
- [Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions](https://mlsanity.com/paper/2507.05257/) (2025)
- [Memorizing Transformers](https://mlsanity.com/paper/2203.08913/) (2022)
- [Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attention](https://mlsanity.com/paper/2404.07143/) (2024)
- [Kimi Linear: An Expressive, Efficient Attention Architecture](https://mlsanity.com/paper/2510.26692/) (2025)
- [Augmenting Language Models with Long-Term Memory](https://mlsanity.com/paper/2306.07174/) (2023)
- [MIRIX: Multi-Agent Memory System for LLM-Based Agents](https://mlsanity.com/paper/2507.07957/) (2025)
- [Context Length Alone Hurts LLM Performance Despite Perfect Retrieval](https://mlsanity.com/paper/2510.05381/) (2025)
- [MemGen: Weaving Generative Latent Memory for Self-Evolving Agents](https://mlsanity.com/paper/2509.24704/) (2025)
- [Hymba: A Hybrid-head Architecture for Small Language Models](https://mlsanity.com/paper/2411.13676/) (2024)

## Built on by (most cited first, 13 of 13)

- [Metis: Memory Foundation Model](https://mlsanity.com/paper/2607.26760/) (2026)
- [A-TMA: Decoupling State-Aware Memory Failures in Long-Term Agent Memory](https://mlsanity.com/paper/2607.01935/) (2026)
- [On the Scaling of PEFT: Towards Million Personal Models of Trillion Parameters](https://mlsanity.com/paper/2606.02437/) (2026)
- [Sigma-Mem: An Online Reliability Memory for LLM-based Multi-Agent Systems](https://mlsanity.com/paper/2607.27958/) (2026)
- [LiveMem: Maintaining Memory State Continuity in Long-Running LLM Inference](https://mlsanity.com/paper/2608.02515/) (2026)
- [Macaron-V1: Towards Open Continual Learning with Self-Improvement and Mixture-of-LoRA](https://mlsanity.com/paper/2608.09819/) (2026)
- [MemBodied: Recurrent Associative Memory for Vision-Language-Action Models](https://mlsanity.com/paper/2609.28256/) (2026)
- [LifeFuse-Mem: Lifecycle-Aware State Fusion Against Temporary Overwriting for Long-Term Memory](https://mlsanity.com/paper/2609.12436/) (2026)
- [PREM: Prefix-Steered Recurrent Memory for Long-Video Understanding](https://mlsanity.com/paper/2609.23601/) (2026)
- [BaRe-Mem: Bayesian Reliability Memory for Robust and Adaptive Agent Consultation](https://mlsanity.com/paper/2609.35551/) (2026)
- [The Optimization Landscape of Learning Compacted Context Models](https://mlsanity.com/paper/2610.05885/) (2026)
- [Towards In-Parameter Memory Augmentation for Large Language Models](https://mlsanity.com/paper/2610.08630/) (2026)
- [DPPM: Dual-Path Parametric Memory for Personalized Language Models](https://mlsanity.com/paper/2610.11776/) (2026)

## Abstract (arXiv)

Large language models increasingly need to accumulate and reuse historical information in long-term assistants and agent systems. Simply expanding the context window is costly and often fails to ensure effective context utilization. We propose δ-mem, a lightweight memory mechanism that augments a frozen full-attention backbone with a compact online state of associative memory. δ-mem compresses past information into a fixed-size state matrix updated by delta-rule learning, and uses its readout to generate low-rank corrections to the backbone's attention computation during generation. With only an 8×8 online memory state, δ-mem improves the average score to 1.10× that of the frozen backbone and 1.15× that of the strongest non-δ-mem memory baseline. It achieves larger gains on memory-heavy benchmarks, reaching 1.31× on MemoryAgentBench and 1.20× on LoCoMo, while largely preserving general capabilities. These results show that effective memory can be realized through a compact online state directly coupled with attention computation, without full fine-tuning, backbone replacement, or explicit context extension.