LongMemEval-S

The 'LongMemEval-S' dataset/benchmark is used to evaluate the performance of memory systems in long-context language model agents by assessing their ability to manage and utilize persistent state across interactions.

Papers using LongMemEval-S (15)

Memory-R1: Enhancing Large Language Model Agents to Manage and Utilize Memories via Reinforcement Learning2025 · 98 citesEpiCache: Episodic KV Cache Management for Long-Term Conversation on Resource-Constrained Environments2025Mnemis: Dual-Route Retrieval on Hierarchical Graphs for Long-Term LLM Memory2026Supersede: Diagnosing and Training the Memory-Update Gap in LLM Agents2026Memory Shot for Long-Term Dialogue2026MemDelta: Controlled Baselines and Hidden Confounds in Agent Memory Evaluation2026MemMachine: A Ground-Truth-Preserving Memory System for Personalized AI Agents2026Learning Query-Aware Budget-Tier Routing for Runtime Agent Memory2026MemForest: An Efficient Agent Memory System with Hierarchical Temporal Indexing2026FlashMemory-DeepSeek-V4: Lightning Index Ultra-Long Context via Lookahead Sparse Attention2026DimMem: Dimensional Structuring for Efficient Long-Term Agent Memory2026MemSkill: Learning and Evolving Memory Skills for Self-Evolving Agents2026Beyond Dialogue Time: Temporal Semantic Memory for Personalized LLM Agents2026LightMem: Lightweight and Efficient Memory-Augmented Generation2025In Prospect and Retrospect: Reflective Memory Management for Long-term Personalized Dialogue Agents2025