GRAVITY: Architecture-Agnostic Structured Anchoring for Long-Horizon Conversational Memory
Not on the map. Why?
Abstract (arXiv)
Long-horizon memory systems increasingly improve how evidence is stored and retrieved, yet the generator must still reason over fragments whose cross-session relationships are implicit. We study generation-time memory organization as a distinct design dimension and introduce GRAVITY (Generation-time Relational Anchoring Via Injected Topological MemorY), a host-independent auxiliary memory layer. GRAVITY consolidates raw dialogue into entity profiles, temporal event traces, and cross-session topic summaries, then retrieves and injects query-relevant records through the prompt interface. Across five heterogeneous memory systems on LongMemEval and LoCoMo, it improves every host--benchmark baseline under two distinct LLM configurations. Controlled analyses separate gains from organizing already available evidence and from consolidating information across the full history. Under a matched LightMem pipeline, the entity--event--topic representation reaches 83.9% on LoCoMo, 3.6% above the strongest of six alternative auxiliary representations. These results show that generation-time structure is a portable complement to existing memory retrieval, while its interaction with host evidence depends on the benchmark and host.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 27 of its 30 references that are in the map. First posted on arXiv 3 May 2026.
Builds on
27 of 30 references.
References · 27
- 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
- 2019Parameter-Efficient Transfer Learning for NLPcited by 4,195
- 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,072
- 2023Generative Agents: Interactive Simulacra of Human Behaviorcited by 3,500
- 2023Retrieval-Augmented Generation for Large Language Models: A Surveycited by 2,185
- 2024From Local to Global: A Graph RAG Approach to Query-Focused Summarizationcited by 1,411
- 2023MemGPT: Towards LLMs as Operating Systemscited by 1,295
- 2025A-MEM: Agentic Memory for LLM Agentscited by 1,000
- 2024Evaluating Very Long-Term Conversational Memory of LLM Agentscited by 800
- 2018MultiWOZ -- A Large-Scale Multi-Domain Wizard-of-Oz Dataset for Task-Oriented Dialogue Modellingcited by 760
- 2025Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memorycited by 722
- 2024LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memorycited by 656
15 more
- 2024A Survey on the Memory Mechanism of Large Language Model based Agentscited by 622
- 2023MemoryBank: Enhancing Large Language Models with Long-Term Memorycited by 539
- 2024RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrievalcited by 519
- 2025Zep: A Temporal Knowledge Graph Architecture for Agent Memorycited by 372
- 2024LightRAG: Simple and Fast Retrieval-Augmented Generationcited by 310
- 2021Beyond Goldfish Memory: Long-Term Open-Domain Conversationcited by 266
- 2025From RAG to Memory: Non-Parametric Continual Learning for Large Language Modelscited by 223
- 2023Augmenting Language Models with Long-Term Memorycited by 186
- 2025LightMem: Lightweight and Efficient Memory-Augmented Generationcited by 163
- 2024CRAG -- Comprehensive RAG Benchmarkcited by 94
- 2025LiCoMemory: Lightweight and Cognitive Agentic Memory for Efficient Long-Term Reasoningcited by 37
- 2024Are Large Language Models a Good Replacement of Taxonomies?cited by 19
- 2026Membox: Weaving Topic Continuity into Long-Range Memory for LLM Agentscited by 6
- 2025KERAG: Knowledge-Enhanced Retrieval-Augmented Generation for Advanced Question Answeringcited by 5
- 2024Creating an AI Observer: Generative Semantic Workspacescited by 3
Built on by
Citing papers per year: 2026: 2.
2 papers.