The Devil behind the mask: An emergent safety vulnerability of Diffusion LLMs Paper • 2507.11097 • Published Jul 15, 2025 • 64
LLMs Learn to Deceive Unintentionally: Emergent Misalignment in Dishonesty from Misaligned Samples to Biased Human-AI Interactions Paper • 2510.08211 • Published Oct 9, 2025 • 22
Your Agent May Misevolve: Emergent Risks in Self-evolving LLM Agents Paper • 2509.26354 • Published Sep 30, 2025 • 17
Demystifying Reasoning Dynamics with Mutual Information: Thinking Tokens are Information Peaks in LLM Reasoning Paper • 2506.02867 • Published Jun 3, 2025 • 2
Test-Time Preference Optimization: On-the-Fly Alignment via Iterative Textual Feedback Paper • 2501.12895 • Published Jan 22, 2025 • 61
