Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages

• URL: http://arxiv.org/abs/2607.06596v1
• Date: Mon, 06 Jul 2026 22:40:07 GMT
• Title: Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Abstract summary: We make the untrusted policy family the controlled axis and decompose the cross-family AUROC into how obvious each family's sabotage is.<n>On code-backdoor transcripts the interaction is positive and survives the dominant confounds.<n>We give a four-step protocol that acts on the gap, with each step a measured result.
• Score: 0.7252027234425333
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Trusted monitoring is a central defense in AI control: a cheaper trusted model scores an untrusted model's actions for sabotage, and the most suspicious are audited or deferred. Such monitors are evaluated against one or two untrusted models, and the accuracy is reported as a property of the monitor. We ask whether it is partly a property of the pairing. We make the untrusted policy family the controlled axis: we fit a monitor on family A's transcripts, apply it to family B, and decompose the cross-family AUROC into how obvious each family's sabotage is, how capable each monitor is, and the residual own-family advantage after both are removed: the interaction. On code-backdoor transcripts the interaction is positive and survives the dominant confounds: +0.172 (95% CI [+0.158, +0.185]) on four open-weight families on a strict leak-free basis, holding across an eight-family matrix spanning four frontier vendors, blatancy-matched controls, a trained transcript-reading monitor up to 34B parameters, and a second code benchmark. The gap is geometric: each lineage's sabotage occupies a different scoring direction, and rotating the monitor's scoring direction toward the target lineage recovers most of the lost detection while a matched random rotation recovers nothing. At a 1% audit budget under defer-to-trusted, an off-lineage monitor catches 19% of attack attempts where an in-lineage monitor catches 41% at the same audit cost, a +21.7-point gap. Control evaluations should report cross-family transfer matrices: a single-pairing accuracy overstates the safety a monitor delivers against a lineage it was not calibrated on. We give a four-step protocol that acts on the gap, with each step a measured result.

Related papers

• Beyond the Transcript: Detecting Covert Co ordination in Latent Multi-Agent Communication [19.840551611758247]
  Language-model agents can communicate through continuous hidden states that are invisible in public transcripts.<n>We introduce Verifiable Latent Alignments (VLA), an activation-aware framework for monitoring and steering these private communication channels.
  arXiv  Detail & Related papers  (2026-08-19T17:43:22Z)
• Beyond Uncertainty: Generalizable Failure Monitoring for Surgical Segmentation under Acquisition Degradation [2.287641923890767]
  Surgical segmentation networks can fail silently under acquisition degradation.<n>We present TCSR-Monitor, a post-hoc failure-monitoring framework that combines confidence with observable shape, temporal-consistency, and image-quality cues.
  arXiv  Detail & Related papers  (2026-08-17T15:57:19Z)
• 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)
• The Distributed Detectability Band Against Marginal-Preserving Attacks [0.0]
  AI-control monitors score individual agent actions to detect misbehavior.<n>Real harm can be distributed across many benign-looking steps, each individually below any per-step alarm.<n>We construct a marginal-preserving, correlation-encoded distributed-sabotage attack.
  arXiv  Detail & Related papers  (2026-06-09T06:04:30Z)
• Stateful Online Monitoring Catches Distributed Agent Attacks [63.00394432922707]
  We develop an online stateful monitor that uses real-time clustering to collect weak suspiciousness signals across many agent transcripts.<n>Our results point toward a new class of safety monitors which reason over groups of users rather than isolated transcripts.
  arXiv  Detail & Related papers  (2026-05-29T17:57:00Z)
• 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)
• 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 Linear Probes: Dynamic Safety Monitoring for Language Models [67.15793594651609]
  Traditional safety monitors require the same amount of compute for every query.<n>We introduce Truncated Polynomials (TPCs), a natural extension of linear probes for dynamic activation monitoring.<n>Our key insight is that TPCs can be trained and evaluated progressively, term-by-term.
  arXiv  Detail & Related papers  (2025-09-30T13:32:59Z)
• CoT Red-Handed: Stress Testing Chain-of-Thought Monitoring [3.6284577335311563]
  Chain-of-Thought (CoT) monitoring improves detection by up to 27 percentage points in scenarios where action-only monitoring fails to reliably identify sabotage.<n>CoT traces can also contain misleading rationalizations that deceive the monitor, reducing performance in more obvious sabotage cases.<n>This hybrid monitor consistently outperforms both CoT and action-only monitors across all tested models and tasks, with detection rates over four times higher than action-only monitoring for subtle deception scenarios.
  arXiv  Detail & Related papers  (2025-05-29T15:47:36Z)
• Lie Detector: Unified Backdoor Detection via Cross-Examination Framework [68.45399098884364]
  We propose a unified backdoor detection framework in the semi-honest setting.<n>Our method achieves superior detection performance, improving accuracy by 5.4%, 1.6%, and 11.9% over SoTA baselines.<n> Notably, it is the first to effectively detect backdoors in multimodal large language models.
  arXiv  Detail & Related papers  (2025-03-21T06:12:06Z)
• Malicious Agent Detection for Robust Multi-Agent Collaborative Perception [52.261231738242266]
  Multi-agent collaborative (MAC) perception is more vulnerable to adversarial attacks than single-agent perception. We propose Malicious Agent Detection (MADE), a reactive defense specific to MAC perception. We conduct comprehensive evaluations on a benchmark 3D dataset V2X-sim and a real-road dataset DAIR-V2X.
  arXiv  Detail & Related papers  (2023-10-18T11:36:42Z)

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.