AgentReview is the first LLM-based simulation framework for peer review that quantifies a 37.1% decision variation attributable to reviewer biases.
Backdoor activation attack: Attack large language models using activation steering for safety-alignment
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
Refusal in language models is mediated by a single direction in residual stream activations that can be erased to disable safety or added to elicit refusal.
Soft prompt distillation with total variation and KL divergence transfers safety behaviors from guard models to on-device LLMs and outperforms LoRA adapters, steering vectors, and direct optimization in safety-usefulness trade-offs with minimal inference cost.
Privacy attacks on LLMs show strong signals for membership inference and backdoors but weaker performance for attribute inference and data extraction, with risks highly dependent on system configuration.
Survey of harmful fine-tuning attacks on LLMs, their variants, defense strategies, mechanical analysis, and evaluation methodologies.
citing papers explorer
-
AgentReview: Exploring Peer Review Dynamics with LLM Agents
AgentReview is the first LLM-based simulation framework for peer review that quantifies a 37.1% decision variation attributable to reviewer biases.
-
Refusal in Language Models Is Mediated by a Single Direction
Refusal in language models is mediated by a single direction in residual stream activations that can be erased to disable safety or added to elicit refusal.
-
Distilling Safe LLM Systems via Soft Prompts for On Device Settings
Soft prompt distillation with total variation and KL divergence transfers safety behaviors from guard models to on-device LLMs and outperforms LoRA adapters, steering vectors, and direct optimization in safety-usefulness trade-offs with minimal inference cost.
-
On the Privacy of LLMs: An Ablation Study
Privacy attacks on LLMs show strong signals for membership inference and backdoors but weaker performance for attribute inference and data extraction, with risks highly dependent on system configuration.
-
Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey
Survey of harmful fine-tuning attacks on LLMs, their variants, defense strategies, mechanical analysis, and evaluation methodologies.