What Happens Inside Agent Memory? Circuit Analysis from Emergence to Diagnosis

• URL: http://arxiv.org/abs/2605.03354v2
• Date: Thu, 07 May 2026 05:56:54 GMT
• Title: What Happens Inside Agent Memory? Circuit Analysis from Emergence to Diagnosis
• Abstract summary: We trace circuits across the Qwen-3 family (0.6B--14B) and two memory frameworks (mem0 and A-MEM)<n>We develop an unsupervised stage-level diagnostic that localizes silent failures to the responsible operation up to 76.2% accuracy.
• Score: 19.29853121538131
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Agent memory failures are silent: an LLM-based agent can produce a fluent response even when it fails to extract, retain, or retrieve the information needed across sessions. The write-manage-read loop describes the external pipeline of these systems but leaves open which internal computations implement each stage. Tracing feature circuits across the Qwen-3 family (0.6B--14B) and two memory frameworks (mem0 and A-MEM), we report two mechanistic findings and one deliverable. First, control is detectable before content: routing circuitry is causally active at 0.6B, while content circuitry produces no detectable signal until 4B, exposing a deployment regime where small models route memory decisions before they can reliably extract or ground the underlying facts. Second, the shared hub is recruited, not created: Write and Read converge on a late-layer hub that already exists in the base model as a context-grounding substrate, and memory framing recruits a memory-specific functional direction on this substrate rather than building one of its own. Both findings transfer across mem0 and A-MEM, indicating that the underlying computations are properties of the base model rather than of any particular interface. Building on this circuit structure, we develop an unsupervised stage-level diagnostic that localizes silent failures to the responsible operation up to 76.2% accuracy, outperforming the strongest supervised baseline by 13 points. Together, these results point to circuit-level signatures as a practical handle for monitoring and structurally-guided design of agent memory.

Related papers

• Automata from Agent Traces: Failure and Next-Step Prediction [1.16693435589939]
  We collapse an entire trace corpus into a single, compact finite-state machine (FSM) that serves as a structural substrate for the behavior of LLM agents.<n>For failure prediction, per-state behavioral features reach held-out AUROC up to 0.94, and an online monitor ranks failing runs above passing ones.
  arXiv  Detail & Related papers  (2026-08-24T17:58:01Z)
• Dual-Grained Agent Memory and Shapley Context Attribution for Multimodal Agentic Learner [55.80248281761638]
  multimodal large language models (MLLMs) deliver impressive perception yet still falter on scientific and mathematical reasoning.<n>We propose DG-Mem, a dual-grained agentic memory framework that augments a frozen MLLM with a non- externally stored memory built once from training-time rollouts and consulted read-only at test time.
  arXiv  Detail & Related papers  (2026-08-24T13:56:01Z)
• Dual-Layer Agentic Memory with Fast Write Routing and Slow Consolidation [27.410703803632718]
  We propose a framework that shifts memory management to the write phase through cost-aware routing and periodic parametric consolidation.<n>A 1.7B/8B cascade prunes up to 68% of redundant external memory while escalating fewer than 50% of inputs.<n>We further show that periodic consolidation successfully internalizes external knowledge, allowing the router to adaptively suppress redundant writes as the model's boundaries evolve.
  arXiv  Detail & Related papers  (2026-08-23T04:40:13Z)
• Harness the Memory: A Holistic Evaluation of Memory Substrates in Memory Agents [97.19033877543522]
  We present a controlled harness evaluation of memory substrates for memory-augmented agents.<n>We instrument 26 performance and efficiency metrics under a unified harness.<n>These findings motivate substrate routing as a necessary component of adaptive agent memory systems.
  arXiv  Detail & Related papers  (2026-08-15T03:29:48Z)
• A Topology-Aware, Memory-Centric Architecture that Separates Root-Cause Derivation from Root-Cause Explanation [0.0]
  We argue that the missing ingredient in autonomous operations is not a better anomaly detector or a larger language model, but operational memory.<n>We present O PS C ORTEX, a working multi-agent prototype that organizes this memory into four tiers and uses it to separate two tasks the field usually conflates: deriving a root cause and explaining it.
  arXiv  Detail & Related papers  (2026-06-18T09:20:20Z)
• eMEM: A Hybrid Spatio-Temporal Memory System For Embodied Agents [1.8047694351309203]
  We present eMEM, a hybrid graph-based memory system for embodied agents operating in physical environments.<n>eMEM fills this gap with a multi-index architecture (ITE for structured storage, hlib for approximate nearest neighbour semantic search, and an R-tree for spatial queries)<n>We also introduce eMEM-Bench v1, a benchmark we construct over ProcTHOR-10K scenes for embodied memory evaluation.
  arXiv  Detail & Related papers  (2026-06-02T09:22:12Z)
• Rethinking Memory as Continuously Evolving Connectivity [63.58181883607843]
  Existing memory-augmented LLM agents treat memory as a static repository with pre-defined representations and fixed retrieval pipelines.<n>We propose FluxMem, a connectivity-evolving memory framework that models memory as a heterogeneous graph and progressively refines its topology through three stages.<n>Across three fundamentally distinct benchmarks including LoCoMo, Mind2Web, and GAIA, FluxMem achieves consistent state-of-the-art performance, demonstrating strong adaptation and generalization in complex agentic environments.
  arXiv  Detail & Related papers  (2026-05-27T17:35:34Z)
• The Attribution Blind Spot: Detecting When Language Models Rely on Memory Rather Than Retrieved Context [22.202409807556517]
  We introduce Computational Reality Monitoring to address the attribution blind spot.<n> CRM operationalizes a principle adapted from cognitive science's reality monitoring framework.<n>It detects whether pretraining exposure leaves a measurable internal trajectory signature.
  arXiv  Detail & Related papers  (2026-05-26T09:48:51Z)
• Architectural Observability Collapse in Transformers [0.0]
  Training preserved an internal decision-quality signal that output confidence does not expose.<n>Confidence controls absorb on average 60.3% of raw probe signal across 14 models in 6 families.<n>Training preserved observability where Llama 3.1 8B collapses at identical 32-layer, 32-head, 4096-hidden shape.
  arXiv  Detail & Related papers  (2026-04-27T02:39:02Z)
• MemEvoBench: Benchmarking Memory MisEvolution in LLM Agents [78.95081012334116]
  Large Language Models (LLMs) with persistent memory enhances interaction continuity and personalization but introduces new safety risks.<n> contaminated or biased memory accumulation can trigger abnormal agent behaviors.<n>MemeEvoBench is the first benchmark evaluating memory safety in LLM agents against adversarial memory injection, noisy tool outputs, and biased feedback.
  arXiv  Detail & Related papers  (2026-04-17T07:29:52Z)
• Route-Induced Density and Stability (RIDE): Controlled Intervention and Mechanism Analysis of Routing-Style Meta Prompts on LLM Internal States [10.639846833695806]
  We quantify internal density via activation sparsity, (C2) domain-keyword attention, and (C3) output stability via predictive entropy and semantic variation.<n>We present RIDE as a diagnostic probe for calibrating routing design and uncertainty estimation.
  arXiv  Detail & Related papers  (2026-03-31T03:19:36Z)
• MemMA: Coordinating the Memory Cycle through Multi-Agent Reasoning and In-Situ Self-Evolution [52.29303869050117]
  Memory-augmented LLM agents maintain external memory banks to support long-horizon interaction.<n>MemMA is a plug-and-play multi-agent framework that coordinates the memory cycle along both the forward and backward paths.
  arXiv  Detail & Related papers  (2026-03-19T10:15:59Z)
• SE-Bench: Benchmarking Self-Evolution with Knowledge Internalization [52.635237306338574]
  We introduce SE-Bench, a diagnostic environment that obfuscates the NumPy library and its API doc into a pseudo-novel package with randomized identifiers.<n>Agents are trained to internalize this package and evaluated on simple coding tasks without access to documentation.<n>Our investigation reveals three insights: (1) the Open-Book Paradox, where training with reference documentation inhibits retention, requiring "Closed-Book Training" to force knowledge compression into weights; (2) the RL Gap, where standard RL fails to internalize new knowledge completely due to PPO clipping and negative gradients; and (3) the viability of Self-Play for internalization, proving models
  arXiv  Detail & Related papers  (2026-02-04T17:58:32Z)
• TraceRouter: Robust Safety for Large Foundation Models via Path-Level Intervention [44.64827167753535]
  harmful semantics act as distributed, cross-layer circuits, rendering localized interventions brittle and detrimental to utility.<n>We propose textbfTrace, a path-level framework that traces and disconnects the causal propagation circuits of illicit semantics.<n>Trace significantly outperforms state-of-the-art baselines, achieving a superior trade-off between adversarial robustness and general utility.
  arXiv  Detail & Related papers  (2026-01-29T15:58:12Z)
• Why Does the LLM Stop Computing: An Empirical Study of User-Reported Failures in Open-Source LLMs [50.075587392477935]
  We conduct the first large-scale empirical study of 705 real-world failures from the open-source DeepSeek, Llama, and Qwen ecosystems.<n>Our analysis reveals a paradigm shift: white-box orchestration relocates the reliability bottleneck from model algorithmic defects to the systemic fragility of the deployment stack.
  arXiv  Detail & Related papers  (2026-01-20T06:42:56Z)
• RePCS: Diagnosing Data Memorization in LLM-Powered Retrieval-Augmented Generation [0.0]
  Models may still rely on memorized training data, bypass the retrieved evidence, and produce contaminated outputs.<n>We introduce Retrieval-Path Contamination Scoring (RePCS), a diagnostic method that detects such behavior without requiring model access or retraining.
  arXiv  Detail & Related papers  (2025-06-18T14:48:19Z)
• Position-aware Automatic Circuit Discovery [59.64762573617173]
  We identify a gap in existing circuit discovery methods, treating model components as equally relevant across input positions.<n>We propose two improvements to incorporate positionality into circuits, even on tasks containing variable-length examples.<n>Our approach enables fully automated discovery of position-sensitive circuits, yielding better trade-offs between circuit size and faithfulness compared to prior work.
  arXiv  Detail & Related papers  (2025-02-07T00:18:20Z)
• Binary DAD-Net: Binarized Driveable Area Detection Network for Autonomous Driving [94.40107679615618]
  This paper proposes a novel binarized driveable area detection network (binary DAD-Net) It uses only binary weights and activations in the encoder, the bottleneck, and the decoder part. It outperforms state-of-the-art semantic segmentation networks on public datasets.
  arXiv  Detail & Related papers  (2020-06-15T07:09:01Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.