Summary The Stages-Oversight benchmark from the Situational Awareness Dataset tests an LLM’s ability to distinguish evaluation prompts from deployment prompts. This capability is a prerequisite for many rogue LLM threat scenarios like sandbagging and alignment faking, where an LLM behaves differently between evaluation and deployment. In my previous post, I...
Summary The Stages-Oversight benchmark from the Situational Awareness Dataset tests whether large language models (LLMs) can distinguish between evaluation prompts (such as benchmark questions) and deployment prompts (real-world user inputs). This ability is crucial for detecting potential alignment risks like sandbagging (strategically underperforming during evaluation) or alignment faking (appearing aligned...
The Stages-Oversight benchmark from the Situational Awareness Dataset tests an LLM’s ability to distinguish evaluation prompts from deployment prompts. This capability is a prerequisite for many rogue LLM threat scenarios like sandbagging and alignment faking, where an LLM behaves differently between evaluation and deployment.
In my previous post, I modified Stages-Oversight to create two new benchmarks:
The Stages-Oversight benchmark from the Situational Awareness Dataset tests whether large language models (LLMs) can distinguish between evaluation prompts (such as benchmark questions) and deployment prompts (real-world user inputs). This ability is crucial for detecting potential alignment risks like sandbagging (strategically underperforming during evaluation) or alignment faking (appearing aligned during testing but pursuing misaligned goals post-deployment).
Surprisingly, most frontier LLMs—including gpt-4o and claude-3.5-sonnet—fail to outperform random chance (50%) on Stages-Oversight, suggesting they do not reliably differentiate evaluation from deployment. I thought this poor performance was unusual, so I created two modified benchmarks to take a closer look at LLM situational awareness: