TL;DR: We introduce CLEVER, a hand-curated benchmark for verified code generation in Lean. It requires full formal specs and proofs. No few-shot method solves all stages, making it a strong testbed for synthesis and formal reasoning.
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-conflicting), and explains why models latch onto shortcuts (simplicity bias, training dynamics). 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- ence stage. In CLEVER, the claim-evidence fusion model and the claim-only model are independently trained to capture the corresponding information.
Clever.musdsoundnik Detail, 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. 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 and is computationally feasible for large neural networks. " This paper introduces a clever incorporation of knowledge graph operation for structured RAG " (Reviewer ifaQ).
Clever.musdsoundnik Detail, " The proposed method is straightforward, intuitive, and easy to implement "; " It is innovative that the paper leverages the structured nature of reasoning paths to filter and refine generated trajectories for model training ... 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 tested this setup on a subset of the failed instances in the one-shot natural language prompt configuration using GPT-4, given its larger context window. We use a clever technique that involves rotating the data within each layer of the model, making it easier to identify and keep only the most important parts for processing. This ensures that the model remains fast and efficient without losing much accuracy.
