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; it evaluates …

2025年9月18日 · This paper introduces CLEVER, a benchmark dataset designed to evaluate LLMs on formally verified code generation. It consists of 161 carefully crafted Lean specifications derived from …

Promoting openness in scientific communication and the peer-review process

2026年2月21日 · This survey on spurious correlations uses the Clever Hans metaphor to motivate the problem, formalizes a group-based setup g=(y,a) with core metrics (worst-group, average-group, bias …

2024年9月25日 · In this paper, we revisit the roles of augmentation strategies and equivariance in improving CL's efficacy. We propose CLeVER (Contrastive Learning Via Equivariant …

579 In this paper, we have proposed a novel counter- factual framework CLEVER for debiasing fact- checking models. Unlike existing works, CLEVER is augmentation-free and mitigates biases on infer- …

2018年2月15日 · Our analysis yields a novel robustness metric called CLEVER, which is short for Cross Lipschitz Extreme Value for nEtwork Robustness. The proposed CLEVER score is attack-agnostic …

2025年5月1日 · 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 …

2025年1月22日 · Leaving the barn door open for Clever Hans: Simple features predict LLM benchmark answers Lorenzo Pacchiardi, Marko Tesic, Lucy G Cheke, Jose Hernandez-Orallo 27 Sept 2024 …

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. We …