Leaving the barn door open for clever hans 05 feb 2025) submitted to iclr 2025 readers We introduce clever, the first curated benchmark for evaluating the generation of specifications and formally verified code in lean The benchmark comprises of 161 programming problems Our analysis yields a novel robustness metric called clever, which is short for cross lipschitz extreme value for network robustness One common approach is training models to refuse unsafe queries, but this strategy can be vulnerable to clever prompts, often referred to as jailbreak attacks, which can trick the ai into providing harmful responses
Our method, stair (safety alignment with introspective reasoning), guides models to think more carefully before responding. Many intellectual endeavors require mathematical problem solving, but this skill remains beyond the capabilities of computers To measure this ability in machine learning models, we introduce math, a new dataset of 12,500 challenging competition mathematics problems While, as we mentioned earlier, there can be thorny “clever hans” issues about humans prompting llms, an automated verifier mechanically backprompting the llm doesn’t suffer from these Membership inference and memorization is a key challenge with diffusion models Mitigating such vulnerabilities is hence an important topic
