SOURCE-LINKED INTELLIGENCE

Stress-Testing Efficient Responsible-AI Evaluation: When Compute Savings Change Benchmark Conclusions

arXiv · AI, language, vision and robotics · article · Aug 31, 2026 · UTC

Efficient evaluation changes the protocol used to support claims about model behavior, yet it is rarely tested whether those claims remain stable after the evaluation itself is made cheaper. We stress-test conclusion robustness in responsible-AI benchmarking by evaluating three dense and mixture-of-experts models on BBQ and BBQ-V under seven conditions spanning batching, quantization, benchmark reduction, and their combinations. Rather than treating preserved aggregate accuracy as sufficient, we compare accuracy, bias severity and prevalence, reasoning quality, subgroup behavior, subset-member

Read original source ↗ Open in workspace

recordType

paper

region

Global