MemEmo: Evaluating Emotion in Memory Systems of Agents

Peng Liu; Zhen Tao; Jihao Zhao; Ding Chen; Yansong Zhang; Cuiping Li; Zhiyu Li; Hong Chen

cs.CLFeb 27, 2026

MemEmo: Evaluating Emotion in Memory Systems of Agents

Authors: Peng Liu, Zhen Tao, Jihao Zhao, Ding Chen, Yansong Zhang, Cuiping Li, Zhiyu Li, Hong Chen

Organizations: School of Information, Renmin University of China, Beijing, China · China Telecom Research Institute · MemTensor (Shanghai) Technology

Abstract

Memory systems address the challenge of context loss in Large Language Model during prolonged interactions. However, compared to human cognition, the efficacy of these systems in processing emotion-related information remains inconclusive. To address this gap, we propose an emotion-enhanced memory evaluation benchmark to assess the performance of mainstream and state-of-the-art memory systems in handling affective information. We developed the Human-Like Memory Emotion (HLME) dataset, which evaluates memory systems across three dimensions: emotional information extraction, emotional memory updating, and emotional memory question answering. Experimental results indicate that none of the evaluated systems achieve robust performance across all three tasks. Our findings provide an objective perspective on the current deficiencies of memory systems in processing emotional memories and suggest a new trajectory for future research and system optimization.

Figures & tables

Figure 1: An example illustrates the lack of emotion processing by the memory system in HCI dialogues.

Category	Benchmark	Year	LTMS	MSI	IR	EA	EO	Personalization
Emotional Interaction Frameworks	DAM-LLM	2025	✓	✓	✓	✓	LLM	Dynamic memory adaptation
	SO-AI	2025	✓	✓	✓	✓	$w/o$	Affective support modeling
	EmoHarbor	2026	$\times$	✓	✓	✓	LLM	Inner-state simulation
	EC2ER	2026	$\times$	$\times$	✓	✓	LLM	Synthetic emotional reasoning
	CoEM	2025	✓	$\times$	✓	✓	LLM	Long-text emotion reasoning
Emotional Benchmarks	Emobench	2024	$\times$	$\times$	$\times$	✓	LLM	General EQ assessment

Table 1: Comprehensive comparison of HLME with existing emotional interaction frameworks and emotional benchmarks in terms of long-term memory support ( LTMS ), multi-session interaction ( MSI ), implicit reasoning ( IR ), emotion analysis ( EA ), Evaluation Object( EO ),and personalization characteristics. Others represent no specific evaluation model, and MS indicate Memory System.

Figure 2: The HLME dataset construction process pipeline.

Figure 3: HLME dataset evaluation scheme.

Model	Setting	EIE			EMU			EQA		Overall
Model	Setting	Acc $\uparrow$	MAE $\downarrow$	F1 ${}_{\text{slot}}\uparrow$	Acc ${}_{\text{update}}\uparrow$	MAE ${}_{\Delta}\downarrow$	MSS $\uparrow$	Acc ${}_{\text{QA}}\uparrow$	F1 ${}_{\text{evidence}}\uparrow$	Score $\uparrow$
Mem0	Med.	0.4534	1.6667	0.0871	0.0011	1.9977	0.7762	0.50	0.6082	0.3574
Letta	Med.	0.8082	0.9942	0.3632	0.3949	1.21	0.9543	0.6265	0.6945	0.5547
MemOS	Med.	0.707	0.8837	0.2049	0.0205	1.9589	0.7132	0.5531	0.7074	0.4378
Mirix	Med.	0.9629	0.1802	0.8774	0.3378	1.3242	0.9794	0.4352	0.5371	0.5952
MemoBase	Med.	0.8872	0.386	0.4229	0.0433	1.9132	0.3687	0.5271	0.662	0.4541

Table 2: Overall emotion-related performance of different memory systems on Medium (Med.) and Long.

\uparrow

\downarrow

indicate higher / lower is better. Best results are in bold , worst results are underlined .

Dataset	System	Add	Search
Medium	Mem0	141.10	1624.33
	Letta	1085.86	4075.66
	MemOS	622.88	1665.64
	Mirix	123.83	4821.08
	MemoBase	3478.41	4895.94
Large	Mem0	356.01	2868.70

Table 3: Memory operation latency on Medium and Large version datasets, The time unit in the experimental results is expressed in seconds. Add represents the time when all the data is added to the memory system, while search represents the total time for multiple searches such as top-10, top-20, and top-50. Best results are in bold , worst results are underlined .

Dataset	Model	EQA (T20)		EQA (T50)
Dataset	Model	Acc ${}_{\text{QA}}$	F1 ${}_{\text{evid.}}$	Acc ${}_{\text{QA}}$	F1 ${}_{\text{evid.}}$
Medium	Mem0	0.505	0.606	0.486	0.597
	Letta	0.620	0.439	0.646	0.716
	MemOS	0.532	0.690	0.563	0.741
	Mirix	0.462	0.562	0.474	0.570
	MemoBase	0.548	0.689	0.568	0.694

Table 4: Emotion QA (EQA) performance of different memory systems on Medium and Large settings. T20 and T50 represent top-20 and top-50.Best results are in bold , worst results are underlined .

Explore similar work

ENPMR-Bench: Benchmarking Proactive Memory Retrieval for Emotional Support Agents
May 26, 2026Xing Fu, Yulin Hu, Mengtong Ji +5Emotional Support Conversation AI Agent Benchmarks
Research Center for Social Computing and Interactive Robotics Harbin Institute of Technology, Heilongjiang, China
EvoMemBench: Benchmarking Agent Memory from a Self-Evolving Perspective
May 18, 2026Yuyao Wang, Zhongjian Zhang, Mo Chi +7LLM Agent Memory LLM Agent Evaluation
Hong Kong University of Science and Technology (Guangzhou) · Beijing University of Posts and Telecommunications · Beijing Institute of Technology +1
MEME: Multi-entity & Evolving Memory Evaluation
May 12, 2026Seokwon Jung, Alexander Rubinstein, Arnas Uselis +2Agent Memory Persistent Memory for Language Models
1KAIST AI · 2Tübingen AI Center, University of Tübingen · 3NAVER AI Lab

May 26, 2026J/K move · Enter open · S save

ENPMR-Bench: Benchmarking Proactive Memory Retrieval for Emotional Support Agents

Xing Fu, Yulin Hu, Mengtong Ji, Haozhen Li+4

Research Center for Social Computing and Interactive Robotics Harbin Institute of Technology, Heilongjiang, China

Emotional Support Conversation AI Agent Benchmarks Memory-Augmented Language Model Agents

Open paper AskPDF arXiv

Memory-augmented language agents are increasingly deployed in affective applications such as emotional support, where understanding and responding to users' latent emotional needs is critical. However, existing research often treats memory as a tool for factual retrieval, overlooking its role in shaping users' emotional experiences. In this work, we introduce ENPMR-Bench, a benchmark for evaluating Emotional Need-aware Proactive Memory Retrieval (ENPMR), a core capability that enables agents to infer users' latent emotional needs and proactively retrieve appropriate memories to support empathetic interaction. Grounded in Maslow's hierarchy of needs, ENPMR-Bench includes over 1,800 memory-augmented dialogues and defines structured mappings between emotional needs and supportive memory types. Experimental results demonstrate that current retrieval paradigms, including both embedding-based and LLM-driven approaches, exhibit substantial deficiencies, with empathy scores significantly lagging behind golden memory conditions. While chain-of-thought prompting improves the alignment between inferred emotional needs and retrieved memories to some extent, a notable performance gap remains. Together, these findings reveal critical limitations in current agents and outline directions for advancing personalized emotional support through need-sensitive memory retrieval.

Abstract

Explore similar work

ENPMR-Bench: Benchmarking Proactive Memory Retrieval for Emotional Support Agents

EvoMemBench: Benchmarking Agent Memory from a Self-Evolving Perspective

MEME: Multi-entity & Evolving Memory Evaluation