Neuromem: A Granular Decomposition of the Streaming Lifecycle in External Memory for LLMs

TL;DR

Neuromem introduces a comprehensive benchmarking framework for streaming external memory modules in large language models, analyzing their lifecycle and performance under dynamic, real-time data ingestion and retrieval scenarios.

Contribution

It provides a scalable testbed that decomposes the memory lifecycle into five dimensions and evaluates various memory strategies on multiple datasets.

Findings

Memory performance degrades as memory size increases.

Time-related queries are the most challenging.

Memory data structure significantly impacts quality and efficiency.

Abstract

Most evaluations of External Memory Module assume a static setting: memory is built offline and queried at a fixed state. In practice, memory is streaming: new facts arrive continuously, insertions interleave with retrievals, and the memory state evolves while the model is serving queries. In this regime, accuracy and cost are governed by the full memory lifecycle, which encompasses the ingestion, maintenance, retrieval, and integration of information into generation. We present Neuromem, a scalable testbed that benchmarks External Memory Modules under an interleaved insertion-and-retrieval protocol and decomposes its lifecycle into five dimensions including memory data structure, normalization strategy, consolidation policy, query formulation strategy, and context integration mechanism. Using three representative datasets LOCOMO, LONGMEMEVAL, and MEMORYAGENTBENCH, Neuromem evaluates interchangeable variants within a shared serving stack, reporting token-level F1 and insertion/retrieval latency. Overall, we observe that performance typically degrades as memory grows across rounds, and time-related queries remain the most challenging category. The memory data structure largely determines the attainable quality frontier, while aggressive compression and generative integration mechanisms mostly shift cost between insertion and retrieval with limited accuracy gain.