Tag: agentic AI safety
Browse our exclusive articles!
Positive Alignment: AI for Human and Ecological Flourishing
Explore Positive Alignment in AI to ensure safety, cooperation, and support for human and ecological flourishing through ethical, context-aware design.
IndustryBench: Benchmarking LLMs for Safe Industrial QA
Discover IndustryBench, a benchmark evaluating large language models on safety and accuracy in industrial procurement QA across multiple languages.
Safety Risks of Malicious Knowledge Editing in AI Models
Explore EditRisk-Bench, a new benchmark assessing safety risks in AI reasoning caused by malicious knowledge editing in large language models.
Do Linear Probes Generalize Better Using Persona Coordinates?
Explore how linear probes with persona coordinates improve generalization in detecting harmful behaviors in language models across datasets.
How Attention Heads Influence Persuasion in LLMs
Discover how a few attention heads reroute large language models, revealing key mechanisms behind LLM persuasion and AI safety.