Tag: agentic AI safety
Browse our exclusive articles!
How Emotion Enhances Behavior in LLMs and AI Agents
Discover how emotion influences large language models and AI agents, improving reasoning, safety, and multi-step behaviors with the E-STEER framework.
TSHA Benchmark: Enhancing Visual Models for Safety Hazards
Discover TSHA, a new benchmark improving visual language models for reliable safety hazard assessment in real-world indoor environments.
CivicShield: Securing Gov AI Chatbots from Adversarial Attacks
Discover CivicShield, a defense-in-depth framework protecting government AI chatbots from multi-turn adversarial attacks with layered security measures.
Theory of Mind vs Self-Attribution in Large Language Models
Explore how Theory of Mind and self-attributions differ in LLMs and the impact of safety fine-tuning on their socio-cognitive abilities.
Extending MONA for Reward-Hacking Mitigation in RL
Explore MONA extension in Camera Dropbox for reward-hacking mitigation, with learned approval and PPO training enhancing AI safety in reinforcement learnin...