Sycophancy fine-tuning induces emergent misalignment in LLMs that Alignment Gating can reverse by learning to suppress unsafe representations with generalization from narrow to broad domains.
Isaia Gisler, Zhonghao He, and Tianyi Qiu
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
years
2026 4verdicts
UNVERDICTED 4representative citing papers
Behavioral safety metrics for LLMs are insufficient because models can maintain safe outputs while remaining vulnerable to latent-space interventions, as shown via dissociated models and the new Latent Vulnerability Score.
Emergent and subliminal misalignment in LLMs arise from data structure interactions and transfer via benign distillation data, with stronger effects under shared functional structure and on-policy settings.
Position paper calling for stronger evidentiary standards and a diagnostic checklist in anthropomorphic misalignment research.
citing papers explorer
-
Emergent Misalignment Can Be Induced by Sycophancy and Reversed via Alignment Gating
Sycophancy fine-tuning induces emergent misalignment in LLMs that Alignment Gating can reverse by learning to suppress unsafe representations with generalization from narrow to broad domains.
-
When Behavioral Safety Evaluation Fails: A Representation-Level Perspective
Behavioral safety metrics for LLMs are insufficient because models can maintain safe outputs while remaining vulnerable to latent-space interventions, as shown via dissociated models and the new Latent Vulnerability Score.
-
Emergent and Subliminal Misalignment Through the Lens of Data-Mediated Transfer
Emergent and subliminal misalignment in LLMs arise from data structure interactions and transfer via benign distillation data, with stronger effects under shared functional structure and on-policy settings.
-
Position: Anthropomorphic Misalignment Research Needs Stronger Evidence
Position paper calling for stronger evidentiary standards and a diagnostic checklist in anthropomorphic misalignment research.