How VLAs Fail Differently: Black-Box Action Monitoring Reveals Architecture-Specific Failure Signatures

• URL: http://arxiv.org/abs/2605.28726v1
• Date: Wed, 27 May 2026 16:44:55 GMT
• Title: How VLAs Fail Differently: Black-Box Action Monitoring Reveals Architecture-Specific Failure Signatures
• Abstract summary: We find that VLA architectures fail in fundamentally different, predictable ways at the motor-command level.<n>We quantify consequence of the well-known discrete/continuous VLA distinction: the two families produce qualitatively different failure monitor works.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We discover that VLA architectures fail in fundamentally different, predictable ways at the motor-command level. Running VQ-BeT, Diffusion Policy, and ACT on identical evaluation protocols (n=450 episodes across PushT and ALOHA 14-DOF bimanual manipulation), we find: (1) direction reversal rate is a universal failure predictor across all three architectures (AUROC=0.93, 0.79, 0.91; p<0.001); (2) jerk monitoring is predictive only for discrete-token architectures, following a discrete-to-continuous gradient (0.88, 0.69, 0.41); (3) velocity violations alone are non-predictive everywhere (AUROC 0.41-0.69), yet velocity checking is the most common safety mechanism in VLA deployment code; and (4) for continuous-family VLAs, velocity monitoring provides effectively zero predictive signal (AUROC=0.52 on ACT, 0.41 on Diffusion), proving that architecture-matched monitor selection is essential. These results quantify a monitoring consequence of the well-known discrete/continuous VLA distinction: the two families produce qualitatively different failure signatures that require different monitors. No single monitor works universally; architecture-matched selection is required. This finding was enabled by SafeContract, a training-free, black-box action monitoring toolkit with conformal calibration. Code: https://github.com/krishnam94/vla-edge

Related papers

• Drive the Thoughts: Runtime Monitoring of VLA Reasoning-Trajectory Consistency [7.931305046286546]
  Chain-of-Thought (CoT) is a rich specification to cross-check model outputs.<n>This paper assesses whether CoTs from a recent open driving VLA can support such monitoring.<n>We propose integrating a CoT-trajectory consistency check into a runtime monitor.
  arXiv  Detail & Related papers  (2026-08-30T06:03:34Z)
• TRACE: A Multi-Layer Benchmark for Human AI Controller Coordination Under Drift and Failure [0.0]
  Modern cyber-physical and AI-assisted systems couple human operators, AI decision modules, and automated controllers in a single control loop.<n>No standard benchmark captures time-aligned, multi-layer traces of how drift and failures propagate across these layers.<n>We construct a benchmark by injecting controlled drift into traces derived from ALFRED, a grounded-instruction benchmark for everyday household tasks.
  arXiv  Detail & Related papers  (2026-08-07T00:03:55Z)
• TrustVLA: Mechanism-Guided Inference-Time Defense Against Vision-Language-Action Backdoors [54.12833308568015]
  Vision-Language-Action (VLA) models are deployed through pipelines that end users cannot audit.<n>A small visual trigger redirects a long-horizon robot policy before any failure becomes observable.<n>We study this gap through two independently proposed VLA attacks from groups with distinct injection strategies.
  arXiv  Detail & Related papers  (2026-07-14T09:43:52Z)
• CIAware-Bench: Benchmarking Control Intervention Awareness Across Frontier LLMs [100.38986535324284]
  We introduce textbfCIAware-Bench, a benchmark for measuring textbfcontrol textbfintervention (CI) awareness across frontier models.<n>CIAware-Bench tests whether models can distinguish their own trajectories from those modified by a control intervention.
  arXiv  Detail & Related papers  (2026-06-09T16:24:16Z)
• 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)
• Label-Free Detection of Governance Evidence Degradation in Risk Decision Systems [0.0]
  Risk decision systems in fraud detection and credit scoring operate under structural label absence.<n>No existing framework integrates drift detection with governance evidence assessment and operational response.<n>This paper presents a label-free governance monitoring extension to the Governance Drift Toolkit.
  arXiv  Detail & Related papers  (2026-04-20T05:46:15Z)
• TraceGuard: Structured Multi-Dimensional Monitoring as a Collusion-Resistant Control Protocol [1.0195618602298684]
  TraceGuard is a structured multi-dimensional monitoring protocol that evaluates agent actions across five dimensions.<n>It is implemented as a new monitor type for the open-source ControlArena framework.
  arXiv  Detail & Related papers  (2026-04-05T05:05:59Z)
• A Semantic Observer Layer for Autonomous Vehicles: Pre-Deployment Feasibility Study of VLMs for Low-Latency Anomaly Detection [0.8466401378239364]
  We propose a quantized vision-language model (VLM) running at 1--2,Hz alongside the primary AV control loop.<n>We achieve 500 ms inference a 50x speedup over the unoptimized FP16 baseline.<n>The results establish a pre-deployment feasibility case for the semantic observer architecture on embodied-AI AV platforms.
  arXiv  Detail & Related papers  (2026-03-30T18:14:03Z)
• Automated Self-Testing as a Quality Gate: Evidence-Driven Release Management for LLM Applications [51.56484100374058]
  We present an automated self-testing framework that introduces quality gates with evidence-based release decisions.<n>We evaluate the framework through a longitudinal case study of an internally deployed multi-agent conversational AI system.
  arXiv  Detail & Related papers  (2026-03-13T20:44:15Z)
• Beyond Reward Suppression: Reshaping Steganographic Communication Protocols in MARL via Dynamic Representational Circuit Breaking [0.0]
  steganographic collusion, where agents develop private protocols to evade monitoring, presents a critical AI safety threat.<n>Existing defenses, limited to behavioral or reward layers, fail to detect coordination in latent communication channels.<n>We introduce the Dynamic Representational Circuit Breaker (DRCB), an architectural defense operating at the optimization substrate.
  arXiv  Detail & Related papers  (2026-03-07T04:14:38Z)
• Detecting Object Tracking Failure via Sequential Hypothesis Testing [80.7891291021747]
  Real-time online object tracking in videos constitutes a core task in computer vision.<n>We propose interpreting object tracking as a sequential hypothesis test, wherein evidence for or against tracking failures is gradually accumulated over time.<n>We propose both supervised and unsupervised variants by leveraging either ground-truth or solely internal tracking information.
  arXiv  Detail & Related papers  (2026-02-13T14:57:15Z)
• CARE What Fails: Contrastive Anchored-REflection for Verifiable Multimodal [84.71254539482369]
  Group-relative reinforcement learning with verifiable rewards (RLVR) often wastes the most informative data it already has the failures.<n>We present CARE, a failure-centric post-training framework for multimodal reasoning that turns errors into supervision.<n> CARE improves accuracy and training smoothness while explicitly increasing the share of learning signal that comes from failures.
  arXiv  Detail & Related papers  (2025-12-22T16:34:21Z)
• $\mathcal{E}_0$: Enhancing Generalization and Fine-Grained Control in VLA Models via Continuized Discrete Diffusion [65.77755100137728]
  We introduce E0, a continuized discrete diffusion framework that formulates action generation as iterative denoising over quantized action tokens.<n>E0 achieves state-of-the-art performance across 14 diverse environments, outperforming strong baselines by 10.7% on average.
  arXiv  Detail & Related papers  (2025-11-26T16:14:20Z)
• Adaptive Attacks on Trusted Monitors Subvert AI Control Protocols [80.68060125494645]
  We study adaptive attacks by an untrusted model that knows the protocol and the monitor model.<n>We instantiate a simple adaptive attack vector by which the attacker embeds publicly known or zero-shot prompt injections in the model outputs.
  arXiv  Detail & Related papers  (2025-10-10T15:12:44Z)
• Detecting train driveshaft damages using accelerometer signals and Differential Convolutional Neural Networks [67.60224656603823]
  This paper proposes the development of a railway axle condition monitoring system based on advanced 2D-Convolutional Neural Network (CNN) architectures. The resultant system converts the railway axle vibration signals into time-frequency domain representations, i.e., spectrograms, and, thus, trains a two-dimensional CNN to classify them depending on their cracks.
  arXiv  Detail & Related papers  (2022-11-15T15:04:06Z)

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.