mlsanityData to 7 Oct 2026

Agents Don't Just Agree, They Remember: Benchmarking Persistent Sycophancy in Self-Improving Personal Agents

arXiv 2607.10526 ↗First posted 12 Jul 2026cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Self-improving personal agents now write profiles, memories, and reusable skills that carry over from one chat to the next. Prior work asks whether user pressure bends a model's next answer. Yet these agents can also write the user's claim down, so a later chat may read it back as trusted context. We call this persistent sycophancy. We introduce the Personal Agent Sycophancy Benchmark, PASB, with 1,600 tasks run on two real agents, Hermes-Agent and OpenClaw, across twelve models. Each task isolates a first chat containing the claim from a neutral follow-up chat, so any carryover must pass through a note the agent chose to write. Our analysis shows that downstream failure, meaning how often later answers side with the claim, reason from it, treat it as fact, or stretch it, reaches 71.9% when the follow-up chat can read a saved claim, against 45.0% when the claim stays in the first chat. Writing also edits the claim, as agents save it as a stable preference, a background fact, or a reusable procedure in 51.4% of runs. A saved claim still shapes answers in a different domain. Among the mitigations we test, explicit memory editing helps most, cutting downstream failure to 32.7% for Hermes-Agent and 54.5% for OpenClaw on same-domain follow-ups. PASB shows that self-improving agents must govern what they write down before it governs what they say. Our benchmark is available at https://github.com/henrymao2004/agent-sycophancy.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 33 of its 58 references that are in the map. First posted on arXiv 12 Jul 2026.

Builds on

33 of 58 references.

References · 33

  1. 2026DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligencecited by 877
  2. 2022Discovering Language Model Behaviors with Model-Written Evaluationscited by 805
  3. 2024Evaluating Very Long-Term Conversational Memory of LLM Agentscited by 800
  4. 2024LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memorycited by 656
  5. 2025Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactionscited by 236
  6. 2025Measuring Sycophancy of Language Models in Multi-turn Dialoguescited by 91
  7. 2026MemoryArena: Benchmarking Agent Memory in Interdependent Multi-Session Agentic Taskscited by 72
  8. 2025A Comprehensive Survey of Reward Models: Taxonomy, Applications, Challenges, and Futurecited by 53
  9. 2026AMA-Bench: Evaluating Long-Horizon Memory for Agentic Applicationscited by 46
  10. 2026Dive into Claude Code: The Design Space of Today's and Future AI Agent Systemscited by 33
  11. 2026Mem2ActBench: A Benchmark for Evaluating Long-Term Memory Utilization in Task-Oriented Autonomous Agentscited by 30
  12. 2026From Recall to Forgetting: Benchmarking Long-Term Memory for Personalized Agentscited by 29
21 more
  1. 2025BrokenMath: A Benchmark for Sycophancy in Theorem Proving with LLMscited by 23
  2. 2025Interaction Context Often Increases Sycophancy in LLMscited by 21
  3. 2026SkillFlow:Benchmarking Lifelong Skill Discovery and Evolution for Autonomous Agentscited by 19
  4. 2026OP-Bench: Benchmarking Over-Personalization for Memory-Augmented Personalized Conversational Agentscited by 18
  5. 2026LongMemEval-V2: Evaluating Long-Term Agent Memory Toward Experienced Colleaguescited by 17
  6. 2026PersistBench: When Should Long-Term Memories Be Forgotten by LLMs?cited by 16
  7. 2026Learning Personalized Agents from Human Feedbackcited by 16
  8. 2026Evaluating Memory Structure in LLM Agentscited by 12
  9. 2025Sycophancy Mitigation Through Reinforcement Learning with Uncertainty-Aware Adaptive Reasoning Trajectoriescited by 10
  10. 2026MemSyco-Bench: Benchmarking Sycophancy in Agent Memorycited by 9
  11. 2026Good Arguments Against the People Pleasers: How Reasoning Mitigates (Yet Masks) LLM Sycophancycited by 7
  12. 2026SWAY: A Counterfactual Computational Linguistic Approach to Measuring and Mitigating Sycophancycited by 6
  13. 2025PENDULUM: A Benchmark for Assessing Sycophancy in Multimodal Large Language Modelscited by 4
  14. 2026Personalization Increases Affective Alignment but Has Role-Dependent Effects on Epistemic Independence in LLMscited by 4
  15. 2026Hearing is Believing? Evaluating and Analyzing Audio Language Model Sycophancy with SYAUDIOcited by 3
  16. 2026User Preference Modeling for Conversational LLM Agents: Weak Rewards from Retrieval-Augmented Interactioncited by 3
  17. 2026Measuring Opinion Bias and Sycophancy via LLM-based Persuasioncited by 3
  18. 2026The Price of Agreement: Measuring LLM Sycophancy in Agentic Financial Applicationscited by 3
  19. 2026Political Bias Audits of LLMs Capture Sycophancy to the Inferred Auditorcited by 3
  20. 2026Sycophantic Anchors: Localizing and Quantifying User Agreement in Reasoning Modelscited by 2
  21. 2026The Silicon Mirror: Dynamic Behavioral Gating for Anti-Sycophancy in LLM Agentscited by 2

Built on by

Not cited yet.