Stern & Nadel (2026)

Drawing on Memory: Dual-Trace Encoding Improves Cross-Session Recall in LLM Agents

arXiv

Find It

DOI

Abstract

LLM agents with persistent memory store information as flat factual records, providing little context for temporal reasoning, change tracking, or cross-session aggregation. Inspired by the drawing effect [3], we introduce dual-trace memory encoding. In this method, each stored fact is paired with a concrete scene trace, a narrative reconstruction of the moment and context in which the information was learned. The agent is forced to commit to specific contextual details during encoding, creating richer, more distinctive memory traces. Using the LongMemEval-S benchmark (4,575 sessions, 100 recall questions), we compare dual-trace encoding against a fact-only control with matched coverage and format over 99 shared questions. Dual-trace achieves 73.7% overall accuracy versus 53.5%, a +20.2 percentage point (pp) gain (95% CI: [+12.1, +29.3], bootstrap p < 0.0001). Gains concentrate in temporal reasoning (+40pp), knowledge-update tracking (+25pp), and multi-session aggregation (+30pp), with no benefit for single-session retrieval, consistent with encoding specificity theory [8]. Token analysis shows dual-trace encoding achieves this gain at no additional cost. We additionally sketch an architectural design for adapting dual-trace encoding to coding agents, with preliminary pilot validation.

Reference

Stern, B., & Nadel, P. (2026). Drawing on Memory: Dual-Trace Encoding Improves Cross-Session Recall in LLM Agents. arXiv preprint arXiv:2604.12948. https://doi.org/10.48550/arxiv.2604.12948

What Rests on This

1 mechanism cites this source. Each finding below is quoted as it appears on that mechanism's own page.

Context-Dependent Retrieval Memory Processes

Measured in AI

For saved memory, a 2026 preprint that stored each fact in an AI agent's long-term memory together with a narrative of the situation in which it was learned raised cross-session recall on the LongMemEval benchmark from 53.5% to 73.7%, an engineering use of encoding specificity with no human comparison