HarmProfile: Characterizing Harmful Distributions in Frontier LLMs
Paper • 2608.14577 • Published • 19
AI & ML interests
Trustworthy ML/AI
Recent Activity
View all activity
hanxunhÂ
authored a
paper 2 days ago
hanxunhÂ
submitted a
paper to Daily Papers 2 days ago
hanxunhÂ
authored 3
papers 10 days ago
When Medical Safety Alignment Fails: A Benchmark for Evaluating LLMs on High-Risk Medical Queries
Paper • 2606.28332 • Published
GeoDetect: Geometric Adversarial Detection for VLPs
Paper • 2607.14737 • Published
Adversarial Attacks for Good: A Survey of Proactive Protection across the Visual Content Lifecycle
Paper • 2608.04314 • Published • 5
hanxunhÂ
submitted a
paper to Daily Papers 10 days ago
hanxunhÂ
authored 4
papers 3 months ago
Density-Aware Translation of Spurious Correlations in Zero-Shot VLMs
Paper • 2606.01710 • Published • 1
AudioMosaic: Contrastive Masked Audio Representation Learning
Paper • 2605.14231 • Published • 3
Safety in Embodied AI: A Survey of Risks, Attacks, and Defenses
Paper • 2605.02900 • Published
Backdoor4Good: Benchmarking Beneficial Uses of Backdoors in LLMs
Paper • 2603.07452 • Published
hanxunhÂ
authored a
paper 5 months ago
hanxunhÂ
authored a
paper 6 months ago
hanxunhÂ
authored 4
papers 7 months ago
Toward Universal and Transferable Jailbreak Attacks on Vision-Language Models
Paper • 2602.01025 • Published
Just Ask: Curious Code Agents Reveal System Prompts in Frontier LLMs
Paper • 2601.21233 • Published
AutoBackdoor: Automating Backdoor Attacks via LLM Agents
Paper • 2511.16709 • Published
BackdoorVLM: A Benchmark for Backdoor Attacks on Vision-Language Models
Paper • 2511.18921 • Published
hanxunhÂ
authored a
paper 12 months ago
hanxunhÂ
authored 2
papers about 1 year ago
hanxunhÂ
authored a
paper over 1 year ago