Safety Benchmark
Noun · AI & Machine Learning
Definitions
A benchmark specifically designed to measure safety-related behavior such as refusal quality, harm avoidance, policy compliance, or robustness to adversarial prompts. Safety benchmarks are useful indicators, though they never cover every real misuse pattern.
In plain English: A benchmark used to test AI safety behavior.
Example: "They complemented the public safety benchmark with internal abuse cases specific to their product domain."