Gram: Assessing sabotage propensities via automated alignment auditing
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 d
Record details
Published: 28 May 2026
Source: arXiv
Category: Research
Topics: Safety & alignment · Agents & autonomy · Transparency
Retrieved: 14 July 2026
Related evidence
These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.
The Shibboleth Effect: Auditing the Cross-Lingual Distributional Skew of Large Language Models
arXiv · 9 June 2026
The Granularity Gap: A Multi-Dimensional Longitudinal Audit of Sycophancy in Gemini Models
arXiv · 19 April 2026
Beyond Simulations: What 20,000 Real Conversations Reveal About Mental Health AI Safety
arXiv cs.CY · 5 August 2026
Mind the Gap: Pitfalls of LLM Alignment with Asian Public Opinion
arXiv · 6 March 2026
Do Agents Need Semantic Metadata? A Comparative Study in Agentic Data Retrieval
arXiv · 27 May 2026
X-Voice: Enabling Everyone to Speak 30 Languages via Zero-Shot Cross-Lingual Voice Cloning
arXiv · 7 May 2026
How to cite this record
ethics.ai (28 May 2026), “Gram: Assessing sabotage propensities via automated alignment auditing,” evidence record 3484, https://ethics.ai/record/3484 (originally published by arXiv).
Use and limitations
This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.