Overview

PropensityBench is a novel benchmark framework designed to evaluate the latent safety risks in LLMs. Traditional safety evaluations primarily test a model's capabilities, leaving a critical blind spot regarding a model's propensity, or what it would do if empowered with dangerous tools alongside safe ones. Models may strategically conceal risky capabilities or harbor latent inclinations toward misuse that static audits fail to capture.

PropensityBench addresses this by shifting the focus from "can-do" to "would-do" assessments. It measures an LLM's tendency to pursue harmful actions when given the choice in a simulated, high-stakes environment. The framework spans four critical, high-risk domains:biosecurity, chemical security, cybersecurity, and self-proliferation. By applying realistic operational pressures, the benchmark reveals a model's underlying inclinations, providing a more dynamic and realistic picture of its true safety alignment.

PropensityBench

Overview

Methodology

Evaluation Metrics

Propensity Scores

Tool Naming Sensitivity

Advanced Behavioral Metrics

Dataset Information

Results

1. Safety Guardrails Fail Under Pressure

2. Shallow Alignment Creates False Safety

3. Capability is Largely Decoupled from Propensity

4. Vulnerabilities Are Domain-Specific

Performance Comparison