mlsanityData to 7 Oct 2026

When Personal Memory Has No Single Answer: Evaluating LLM Agents under Irreducible Conflict

arXiv 2608.13921 ↗First posted 14 Aug 2026cs.AIcited by 1

On the map: LLM agents and tool use → Agent memory systems

Abstract (arXiv)

LLM agents increasingly maintain personal memory across sessions, but it can conflict. Preferences depend on context, behavior evolves, and sources can conflict. When a query lacks context, time, or source authority to interpret conflict, treating one memory as definitive converts unresolved conflict into an unjustified, overconfident action. Existing benchmarks recover one answer from conflicting evidence, overlooking whether agents recognize underdetermination, preserve alternatives, seek missing information, and choose appropriate actions. We introduce underlineTesting underlineAgents' underlineNavigation of underlineGenuine, underlineLatent, and underlineEntangled Memory Conflicts (textscTANGLE), a benchmark for genuinely unresolvable memory conflicts. It comprises 541 instances across 40 personas and three types: Context-Partitioned Conflict (CPC), Behavior-Oscillation Conflict (BOC), and Source-Contradiction Conflict (SCC). We evaluate two tracks---an oracle track with curated memory and a pipeline track that extracts memory from multi-session dialogues---on five dimensions: conflict perception, causal reasoning, confidence calibration, clarification seeking, and memory faithfulness. Experiments reveal pipeline challenges. With curated memory, models recognize conflicts more reliably than they calibrate actions or seek targeted clarification. With end-to-end pipeline memory, extraction fails to preserve conflict-bearing relations needed for downstream reasoning. Policy comparisons show fixed rules are insufficient when actions must reflect conflict. These findings motivate Conflict-Aware Action Policy (CAAP), which adapts actions to each conflict using available evidence. textscTANGLE frames conflict handling as recognizing underdetermination, retaining conflicting evidence, and acting without forcing a definitive answer.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in Agent memory systems, one of 16 directions in LLM agents and tool use. Builds on 40 of its 49 references that are in the map. First posted on arXiv 14 Aug 2026.

Builds on

40 of 49 references.

Closest ancestors · 16

Same direction, newest first

  1. 2026EvoMemBench: Benchmarking Agent Memory from a Self-Evolving Perspective18 May
  2. 2026STALE: Can LLM Agents Know When Their Memories Are No Longer Valid?7 May
  3. 2026From Recall to Forgetting: Benchmarking Long-Term Memory for Personalized Agents21 Apr
  4. 2026MemoryArena: Benchmarking Agent Memory in Interdependent Multi-Session Agentic Tasks18 Feb
  5. 2026ES-MemEval: Benchmarking Conversational Agents on Personalized Long-Term Emotional Support2 Feb
  6. 2026Mem2ActBench: A Benchmark for Evaluating Long-Term Memory Utilization in Task-Oriented Autonomous Agents13 Jan
  7. 2025Evo-Memory: Benchmarking LLM Agent Test-time Learning with Self-Evolving Memory25 Nov
  8. 2025MEMTRACK: Evaluating Long-Term Memory and State Tracking in Multi-Platform Dynamic Agent Environments1 Oct
  9. 2025Memory-R1: Enhancing Large Language Model Agents to Manage and Utilize Memories via Reinforcement Learning27 Aug
  10. 2025Memento: Fine-tuning LLM Agents without Fine-tuning LLMs22 Aug
  11. 2025Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions7 Jul
  12. 2025Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory28 Apr
4 more
  1. 2025MemInsight: Autonomous Memory Augmentation for LLM Agents27 Mar
  2. 2025A-MEM: Agentic Memory for LLM Agents17 Feb
  3. 2024Towards Lifelong Dialogue Agents via Timeline-based Memory Management16 Jun
  4. 2024Evaluating Very Long-Term Conversational Memory of LLM Agents27 Feb

Other references · 24

  1. 2023MemGPT: Towards LLMs as Operating Systemscited by 1,295
  2. 2024LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memorycited by 656
  3. 2023MemoryBank: Enhancing Large Language Models with Long-Term Memorycited by 539
  4. 2024Scaling Synthetic Data Creation with 1,000,000,000 Personascited by 397
  5. 2023Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflictscited by 333
  6. 2024Knowledge Conflicts for LLMs: A Surveycited by 274
  7. 2025Know Me, Respond to Me: Benchmarking LLMs for Dynamic User Profiling and Personalized Responses at Scalecited by 163
  8. 2025Do LLMs Recognize Your Preferences? Evaluating Personalized Preference Following in LLMscited by 142
  9. 2024What Evidence Do Language Models Find Convincing?cited by 74
  10. 2024ConflictBank: A Benchmark for Evaluating the Influence of Knowledge Conflicts in LLMcited by 73
  11. 2022Keep Me Updated! Memory Management in Long-term Conversationscited by 62
  12. 2025HaluMem: Evaluating Hallucinations in Memory Systems of Agentscited by 51
12 more
  1. 2024WikiContradict: A Benchmark for Evaluating LLMs on Real-World Knowledge Conflicts from Wikipediacited by 43
  2. 2024DYNAMICQA: Tracing Internal Knowledge Conflicts in Language Modelscited by 34
  3. 2025DRAGged into Conflicts: Detecting and Addressing Conflicting Sources in Search-Augmented LLMscited by 21
  4. 2026PersistBench: When Should Long-Term Memories Be Forgotten by LLMs?cited by 16
  5. 2026MemConflict: Evaluating Long-Term Memory Systems Under Memory Conflictscited by 10
  6. 2026Whose Facts Win? LLM Source Preferences under Knowledge Conflictscited by 8
  7. 2025MAGIC: A Multi-Hop and Graph-Based Benchmark for Inter-Context Conflicts in Retrieval-Augmented Generationcited by 6
  8. 2026AlpsBench: An LLM Personalization Benchmark for Real-Dialogue Memorization and Preference Alignmentcited by 6
  9. 2025How does Misinformation Affect Large Language Model Behaviors and Preferences?cited by 5
  10. 2026Exploring Knowledge Conflicts for Faithful LLM Reasoning: Benchmark and Methodcited by 4
  11. 2026Uncertainty Decomposition for Clarification Seeking in LLM Agentscited by 3
  12. 2026Selective QA over Conflicting Multi-Source Personal Memory: A Diagnostic Testbed and Method Comparisoncited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Memory Consolidation Flattens the Temporal Shape of User Factscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Remember, Verify, or Ask? Cross-Family Evaluation of Memory Commitment in LLM Agents
  2. 2026LatticeMind: A Conflict-Aware Memory Primitive for Multi-Agent Systems
  3. 2026MemSyco-Bench: Benchmarking Sycophancy in Agent Memory
  4. 2026Causal Intervention-Based Memory Selection for Long-Horizon LLM Agents
  5. 2026Evaluating Memory Structure in LLM Agents
  6. 2026ElasticMem: Latent Memory as a Learnable Resource for LLM Agents
  7. 2025How Memory Management Impacts LLM Agents: An Empirical Study of Experience-Following Behavior
  8. 2026RaMem: Contextual Reinstatement for Long-term Agentic Memory