Ensemble Monitoring for AI Control: Diverse Signals Outweigh More Compute
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917507423535104 |
|---|---|
| author | Koran, Eugene Yun, Yejun Tetef, Samantha Arnav, Benjamin Bernabeu-Pérez, Pablo |
| author_facet | Koran, Eugene Yun, Yejun Tetef, Samantha Arnav, Benjamin Bernabeu-Pérez, Pablo |
| contents | As AI systems are increasingly deployed in autonomous agentic settings at scale, it is important to ensure the actions they take are safe and aligned with user intent. Monitoring agent actions is a key safety mechanism, yet reliable monitors remain difficult to build and the scale of these systems makes human oversight impractical. We show that combining signals from diverse monitors into an ensemble improves detection of misaligned actions. We build 12 GPT-4.1-Mini monitors using both prompting and fine-tuning strategies. We evaluate them on coding tasks where candidate solutions pass standard tests but fail on adversarial inputs. In this setting, diverse ensembles outperform both individual monitors and homogeneous ensembles. Our best 3-monitor ensemble achieves 2.4x greater detection performance gain compared to an ensemble composed of three identical monitors, with the same ensemble performing strongly on an independent dataset. We contend that these results show that diversity - not scale - drives gains. The best ensembles combine strong individual performance with low correlation between monitors. Furthermore, fine-tuned monitors appear in every top-performing ensemble and maintain this advantage on out-of-distribution attack types, suggesting that fine-tuning enables detection capabilities that prompting alone does not elicit. These results support ensemble monitoring as a practical AI control strategy for safety gains at reasonable inference costs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_15377 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Ensemble Monitoring for AI Control: Diverse Signals Outweigh More Compute Koran, Eugene Yun, Yejun Tetef, Samantha Arnav, Benjamin Bernabeu-Pérez, Pablo Artificial Intelligence As AI systems are increasingly deployed in autonomous agentic settings at scale, it is important to ensure the actions they take are safe and aligned with user intent. Monitoring agent actions is a key safety mechanism, yet reliable monitors remain difficult to build and the scale of these systems makes human oversight impractical. We show that combining signals from diverse monitors into an ensemble improves detection of misaligned actions. We build 12 GPT-4.1-Mini monitors using both prompting and fine-tuning strategies. We evaluate them on coding tasks where candidate solutions pass standard tests but fail on adversarial inputs. In this setting, diverse ensembles outperform both individual monitors and homogeneous ensembles. Our best 3-monitor ensemble achieves 2.4x greater detection performance gain compared to an ensemble composed of three identical monitors, with the same ensemble performing strongly on an independent dataset. We contend that these results show that diversity - not scale - drives gains. The best ensembles combine strong individual performance with low correlation between monitors. Furthermore, fine-tuned monitors appear in every top-performing ensemble and maintain this advantage on out-of-distribution attack types, suggesting that fine-tuning enables detection capabilities that prompting alone does not elicit. These results support ensemble monitoring as a practical AI control strategy for safety gains at reasonable inference costs. |
| title | Ensemble Monitoring for AI Control: Diverse Signals Outweigh More Compute |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2605.15377 |