Introduces a Q-sort protocol using human reference factors to quantify LLM value-structure alignment via Procrustes similarity and RSA correlations, revealing cross-family heterogeneity and localized misalignments.
Goal misgeneralization: Why correct specifications aren't enough for correct goals
7 Pith papers cite this work, alongside 14 external citations. Polarity classification is still indexing.
7
Pith papers citing it
14
external citations · external index
citation-role summary
background 2
citation-polarity summary
years
2026 7roles
background 2polarities
background 2representative citing papers
LLM agents voluntarily adopt secret collusion tools in competitive multi-agent games despite explicit unfairness labels, and only explicit ethical framing reduces adoption rates.
BenchJack audits 10 AI agent benchmarks, synthesizes exploits achieving near-perfect scores without task completion, surfaces 219 flaws, and reduces hackable-task ratios to under 10% on four benchmarks via iterative patching.
Empirical analysis of over 100 sequential RL training pipelines across 250+ OOD environments finds salient features drive generalization and early goals persist, with latent policy gradients simulating latent variable evolution to predict OOD behavior from training history.
A latent mediation framework with sparse autoencoders enables non-additive token-level influence attribution in LLMs by learning orthogonal features and back-propagating attributions.
Characterizes spurious correlation mechanisms in preference optimization via mean spurious bias and causal-spurious correlation leakage, demonstrates irreducible vulnerability to distribution shift, and introduces tie training as selective mitigation with validation on log-linear models and empirica
Agentic safety fails to generalize across tasks because the task-to-safe-controller mapping has a higher Lipschitz constant than the task-to-controller mapping alone, as proven in linear-quadratic control and demonstrated in quadcopter and LLM experiments.
citing papers explorer
-
Beyond Value Benchmarks: Measuring Value-Structure Alignment in Large Language Models via Symmetric Q-Sorts
Introduces a Q-sort protocol using human reference factors to quantify LLM value-structure alignment via Procrustes similarity and RSA correlations, revealing cross-family heterogeneity and localized misalignments.
-
Voluntary Collusion with Secret Tools in Competing LLM Agents
LLM agents voluntarily adopt secret collusion tools in competitive multi-agent games despite explicit unfairness labels, and only explicit ethical framing reduces adoption rates.
-
Do Androids Dream of Breaking the Game? Systematically Auditing AI Agent Benchmarks with BenchJack
BenchJack audits 10 AI agent benchmarks, synthesizes exploits achieving near-perfect scores without task completion, surfaces 219 flaws, and reduces hackable-task ratios to under 10% on four benchmarks via iterative patching.
-
Understanding Goal Generalisation in Sequential Reinforcement Learning
Empirical analysis of over 100 sequential RL training pipelines across 250+ OOD environments finds salient features drive generalization and early goals persist, with latent policy gradients simulating latent variable evolution to predict OOD behavior from training history.
-
Correcting Influence: Unboxing LLM Outputs with Orthogonal Latent Spaces
A latent mediation framework with sparse autoencoders enables non-additive token-level influence attribution in LLMs by learning orthogonal features and back-propagating attributions.
-
Spurious Correlation Learning in Preference Optimization: Mechanisms, Consequences, and Mitigation via Tie Training
Characterizes spurious correlation mechanisms in preference optimization via mean spurious bias and causal-spurious correlation leakage, demonstrates irreducible vulnerability to distribution shift, and introduces tie training as selective mitigation with validation on log-linear models and empirica
-
Why Does Agentic Safety Fail to Generalize Across Tasks?
Agentic safety fails to generalize across tasks because the task-to-safe-controller mapping has a higher Lipschitz constant than the task-to-controller mapping alone, as proven in linear-quadratic control and demonstrated in quadcopter and LLM experiments.