Honest Lying: Understanding Memory Confabulation in Reflexive Agents
Summary
This paper identifies memory confabulation in Reflexion-style agents, where agents store incorrect task interpretations and persist in errors across environment resets. The authors introduce the Reflection Repetition Rate (RRR) metric to detect this and propose a mitigation that replaces open-ended self-diagnosis with programmatic failure signal extraction.
View Cached Full Text
Cached at: 06/09/26, 08:41 AM
Paper page - Honest Lying: Understanding Memory Confabulation in Reflexive Agents
Source: https://huggingface.co/papers/2605.29463
Abstract
Agents relying on self-generated reflections can store confident but incorrect task interpretations, leading to persistent errors despite environment resets, which is identified through a new metric called Reflection Repetition Rate.
Reflexion-style agentsrely onself-generated reflectionsas memory, implicitly assuming that agents can accurately diagnose their own failures. We show that this assumption can fail systematically: across ALFWorld and HumanEval, agents store confident but incorrect interpretations of the task and continue acting on them across trials, even though the environment resets to the correct task each time. We call this failure modememory confabulationand introduce theReflection Repetition Rate(RRR), a log-based metric that detects repeated reliance on incorrect reflective content. Using RRR, we identify 16 frozen environments in ALFWorld, where 0 of 121 reflections mention the correct target object, and 4 analogous cases in HumanEval. Our mitigation replaces open-ended self-diagnosis with programmatic extraction oftrajectory-level failure signals, increasing correct object mention from 0% to 86%, reducing RRR from 0.64 to 0.10, and solving 3 of 16 frozen ALFWorld environments, suggesting that reflective memory can reinforce false beliefs rather than correct them.
View arXiv pageView PDFAdd to collection
Get this paper in your agent:
hf papers read 2605\.29463
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2605.29463 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2605.29463 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2605.29463 in a Space README.md to link it from this page.
Collections including this paper0
No Collection including this paper
Add this paper to acollectionto link it from this page.
Similar Articles
Staying In Character: Perspective-Bounded Memory For Book-Based Role-Playing Agents
This paper proposes ReverieMem, a three-layer memory architecture for book-based LLM role-playing agents that prevents factual overreach and stylistic monotony. It also introduces the KBF-QA benchmark and achieves significant improvements in knowledge boundary fidelity and narrative quality.
An agent remembering everything sounds useful until it remembers the wrong crap
The author critiques the idea of agents remembering everything and introduces TrueMemory, a system that converts memories into trait claims with confidence and evidence to better calibrate agent behavior.
What I learned trying to make agent memory survive more than one session
The article reflects on the complexities of AI agent memory beyond simple storage, highlighting challenges such as determining truthfulness, priority changes, distinguishing decisions from noise, and appropriate timing for surfacing context.
The Misattribution Gap: When Memory Poisoning Looks Like Model Failure in Agentic AI Systems
This paper identifies a structural failure in multi-agent AI pipelines where memory-layer attacks can be misattributed as model misalignment, formalizing Semantic Norm Drift (SND) and proposing Counterfactual Composition Testing and Memory-Persistent Information-Flow Control as defenses.
@omarsar0: // The Memory Curse in LLM Agents // (bookmark it) Long histories apparently degrades agents as they become increasingl…
This research paper identifies the 'memory curse' in LLM agents, demonstrating that expanded context windows systematically degrade cooperative behavior in multi-agent social dilemmas by eroding forward-looking intent. The authors show that targeted fine-tuning, synthetic memory sanitization, and reducing explicit Chain-of-Thought reasoning can effectively mitigate this behavioral decay.