OSCBench: Benchmarking Object State Change in Text-to-Video
Discover OSCBench, a benchmark evaluating object state change in text-to-video models, highlighting gaps in action understanding and temporal consistency.
Jailbreak Scaling Laws in Large Language Models Explained
Explore how adversarial attacks cause polynomial to exponential growth in jailbreak success rates of large language models, impacting AI safety.
LLM Robustness to Chain-of-Thought Perturbations Explained
Explore how large language models handle chain-of-thought perturbations and improve reasoning accuracy across diverse error types.
Social-JEPA: Emergent Geometric Isomorphism in AI Models
Discover how Social-JEPA enables interoperable decentralized vision systems via emergent geometric isomorphism for efficient AI collaboration.
Differential Privacy Impact on Medical Image Analysis
Explore how differential privacy affects medical image analysis using DP-RGMI to optimize privacy and model performance in healthcare AI.