The Clever Hans Mirage: A Comprehensive Survey on Spurious. . . 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)
Withdraw - OpenReview Review: The paper is a theoretical analysis of adversarial robustness of Graph Laplacian based semi-supervised learning Pros: The paper addresses an important problem of adversarial robustness by exploiting unlabeled data It studies the theoretical robustness of graph Laplacian approach Cons: The attack model is not very well motivated and connected to some realistic situations The
Evaluating the Robustness of Neural Networks: An Extreme Value. . . 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
CLEVER: A Curated Benchmark for Formally Verified Code Generation 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
Clever: A Curated Benchmark for Formally Verified Code Generation 1 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 both formal speci- fication generation and implementation synthesis from natural language, requiring formal correctness proofs for both
Ignore Previous Prompt: Attack Techniques For Language Models Abstract Transformer-based large language models (LLMs) provide a powerful foundation for natural language tasks in large-scale customer-facing applications However, studies that explore their vulnerabilities emerging from malicious user interaction are scarce By proposing PROMPTINJECT, a prosaic alignment framework for mask-based iterative adversarial prompt composition, we examine how GPT
On the Planning Abilities of Large Language Models : A Critical . . . 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
STAIR: Improving Safety Alignment with Introspective Reasoning 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
LEARNING TO OPTIMIZE - OpenReview This requires careful analysis of existing optimization algorithms and clever solutions to overcome their weaknesses; thus, doing so is both laborious and time-consuming