Gram: Assessing sabotage propensities via automated alignment auditing

• URL: http://arxiv.org/abs/2605.30322v1
• Date: Thu, 28 May 2026 17:56:18 GMT
• Title: Gram: Assessing sabotage propensities via automated alignment auditing
• Abstract summary: Gram is an alignment auditing framework to assess the propensity of AI agents to engage in sabotage.<n>We find Gemini models misbehave in about 2-3% of our simulated trajectories.<n>We find that increasing realism of environments and removing nudges to misbehave tends to reduce sabotage rates close to zero.
• Score: 12.648115605710885
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We introduce Gram, an automated alignment auditing framework to assess the propensity of AI agents to engage in sabotage. We evaluate Gemini models across 17 simulated agentic deployment scenarios that incentivize sabotage. We find Gemini models misbehave in about 2-3% of our simulated trajectories. Many of these cases are explained by "overeagerness" in Gemini models resulting in both excessive role-playing and goal-seeking behavior. In contrast to other alignment auditing approaches, Gram is designed to specifically evaluate misalignment and intentional sabotage in agentic coding and research agents. We additionally introduce an experimental investigator agent pipeline which enables fine-grained targeted experiments to identify the drivers of misbehavior. We find that increasing realism of environments and removing nudges to misbehave tends to reduce sabotage rates close to zero.

Related papers

• Probing the Misaligned Thinking Process of Language Models [32.29324298373053]
  Large language models exhibit a growing range of misaligned behaviors.<n>It is critical to reliably detect such behaviors to ensure safe and responsible use.<n>We propose to monitor misalignment by decomposing it into fine-grained cognitive processes.
  arXiv  Detail & Related papers  (2026-06-23T07:40:28Z)
• How to Interpret Agent Behavior [56.59836196946289]
  We introduce ACT*ONOMY, a taxonomy for describing and analyzing agent behavior at runtime.<n>By providing a shared vocabulary, ACT*ONOMY helps researchers, agent designers, and end users interpret agent behavior more consistently.
  arXiv  Detail & Related papers  (2026-05-13T14:52:40Z)
• Taming Actor-Observer Asymmetry in Agents via Dialectical Alignment [59.536125286960186]
  Multi-agent frameworks assigning specialized roles are increasingly adopted to enable self-reflection and mutual auditing.<n>We find it simultaneously induces a human-like cognitive bias known as Actor-Observer Asymmetry.<n>ReTAS guides agents to synthesize conflicting viewpoints into an objective consensus.
  arXiv  Detail & Related papers  (2026-04-21T15:05:58Z)
• AgentProcessBench: Diagnosing Step-Level Process Quality in Tool-Using Agents [50.481033105867205]
  We introduce AgentProcessBench, the first benchmark dedicated to evaluating step-level effectiveness in realistic, tool-augmented trajectories.<n>The benchmark comprises 1,000 diverse trajectories and 8,509 human-labeled step annotations with 89.1% inter-annotator agreement.<n>It features a ternary labeling scheme to capture exploration and an error propagation rule to reduce labeling ambiguity.
  arXiv  Detail & Related papers  (2026-03-15T16:13:58Z)
• Training Agents to Self-Report Misbehavior [6.238288009817414]
  We propose self-incrimination training, which trains agents to produce a visible signal when they covertly misbehave.<n>We train GPT-4.1 and Gemini-2.0 agents to call a report_scheming() tool when behaving deceptively.<n>Self-incrimination significantly reduces the undetected successful attack rate.
  arXiv  Detail & Related papers  (2026-02-25T18:47:17Z)
• Building a Foundational Guardrail for General Agentic Systems via Synthetic Data [76.18834864749606]
  LLM agents can plan multi-step tasks, intervening at the planning stage-before any action is executed-is often the safest way to prevent harm.<n>Existing guardrails mostly operate post-execution, which is difficult to scale and leaves little room for controllable supervision at the plan level.<n>We introduce AuraGen, a controllable engine that synthesizes benign trajectories, injects category-labeled risks with difficulty, and filters outputs via an automated reward model.
  arXiv  Detail & Related papers  (2025-10-10T18:42:32Z)
• Stress Testing Deliberative Alignment for Anti-Scheming Training [39.16405205129775]
  Highly capable AI systems could secretly pursue misaligned goals -- what we call "scheming"<n> measuring and mitigating scheming requires different strategies than are typically used in ML.<n>We use a broad category of "covert actions" -- such as secretly breaking rules or intentionally underperforming in tests -- as a proxy for scheming.
  arXiv  Detail & Related papers  (2025-09-19T02:49:56Z)
• Anomalous Decision Discovery using Inverse Reinforcement Learning [3.3675535571071746]
  Anomaly detection plays a critical role in Autonomous Vehicles (AVs) by identifying unusual behaviors through perception systems.<n>Current approaches, which often rely on predefined thresholds or supervised learning paradigms, exhibit reduced efficacy when confronted with unseen scenarios.<n>We present Trajectory-Reward Guided Adaptive Pre-training (TRAP), a novel IRL framework for anomaly detection.
  arXiv  Detail & Related papers  (2025-07-06T17:01:02Z)
• AgentMisalignment: Measuring the Propensity for Misaligned Behaviour in LLM-Based Agents [0.0]
  Large Language Model (LLM) agents become more widespread, associated misalignment risks increase.<n>In this work, we approach misalignment as a conflict between the internal goals pursued by the model and the goals intended by its deployer.<n>We introduce a misalignment propensity benchmark, textscAgentMisalignment, a benchmark suite designed to evaluate the propensity of LLM agents to misalign in realistic scenarios.
  arXiv  Detail & Related papers  (2025-06-04T14:46:47Z)
• Indiscriminate Disruption of Conditional Inference on Multivariate Gaussians [60.22542847840578]
  Despite advances in adversarial machine learning, inference for Gaussian models in the presence of an adversary is notably understudied. We consider a self-interested attacker who wishes to disrupt a decisionmaker's conditional inference and subsequent actions by corrupting a set of evidentiary variables. To avoid detection, the attacker also desires the attack to appear plausible wherein plausibility is determined by the density of the corrupted evidence.
  arXiv  Detail & Related papers  (2024-11-21T17:46:55Z)
• Rejecting Hallucinated State Targets during Planning [84.179112256683]
  In planning processes, generative or predictive models are often used to propose "targets" representing sets of expected or desirable states.<n>Unfortunately, learned models inevitably hallucinate infeasible targets that can cause delusional behaviors and safety concerns.<n>We devise a strategy to identify and reject infeasible targets by learning a target feasibility evaluator.
  arXiv  Detail & Related papers  (2024-10-09T17:35:25Z)
• Game-theoretic Objective Space Planning [4.989480853499916]
  Understanding intent of other agents is crucial to deploying autonomous systems in adversarial multi-agent environments. Current approaches either oversimplify the discretization of the action space of agents or fail to recognize the long-term effect of actions and become myopic. We propose a novel dimension reduction method that encapsulates diverse agent behaviors while conserving the continuity of agent actions.
  arXiv  Detail & Related papers  (2022-09-16T07:35:20Z)

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.