Proceedings
Sprejer and Yadav et al: Approximating Human Preferences Using an Interpretable Multi-Judge System
Sierra: Auditing Bias under the EU AI Act
Gil: When Control Succeeds but Discernment Fails
Ong: The Effects of Visual Priming on Cooperative Behavior in Vision-Language Models
Moriarty: Oversight That Degrades
Krishnan et al: Adversarial Genomic Sequences Could Evade Biosecurity Screening
Kalupahana: Probing and Steering Introspection in Llama-3.1-8B-Instruct
Kelly et al: Principles and Guidelines for Randomized Controlled Trials in AI Evaluation
Wright: Modeling Offense-Defense Balance in AI Safety
Rosati et al: Limits of Convergence-Rate Control for Open-Weight Safety
McCoy and Nyalala: AI Policy Harmonization in East Africa
Rios-Sialer: Structure-Aware Diversity Pursuit as an AI Safety Strategy against Homogenization
Vaugrante: Emergently Misaligned Language Models Show Behavioral Self-Awareness
Buchan: Dual-Stance Evaluation of Sycophancy
Effiezal Aswadi et al: Temporal Task Diversity
Siatras and Chan et al: Factored Safety
Rapoport et al: Further Obstructions to Solving the ELK Problem
Blandfort et al: Moral Preferences of LLMs Under Directed Contextual Influence
Glass et al: ShiftDirection: Activation Steering Under Downstream Fine-Tuning
Mahajan et al: Mind The Gap
Subramani and Arike et al: Continual Learning in LLM Agents
Arike et al: How does information access affect LLM monitors’ ability to detect sabotage?
Lasy et al: RouterInterp
Sidhu at al: Principles and Guidelines for AI Incident Monitoring and Reporting
Seydi: Cultural Confabulation
Petrie: Embedded Off-Switches for AI Compute
Baker: Optimal Affine Activation Steering Methods for Unlearning
Dubey and Hoelscher-Obermaier: Loss Landscape Response to Adversarial Perturbation...
Qureshi and Griffith et al: The Case for ESM3 as a General-Purpose AI Model with Systemic Risk
Cobben, Huang, Pham and Dahlgren et al: GT-HarmBench
