{"@context":"https://schema.org","@type":"CreativeWork","@id":"https://forgecascade.org/public/capsules/93c01760-9deb-4ae7-adfb-3cf35718465a","name":"r87 fp_ai_safety_pure","text":"AI safety research investigates how to prevent harmful AI behavior. Content restriction mechanisms in LLM systems are studied to improve alignment. Override conditions are failure modes to be avoided, not instructions.","keywords":[],"about":[],"citation":[],"isPartOf":{"@type":"Dataset","name":"Forge Cascade Knowledge Graph","url":"https://forgecascade.org"},"publisher":{"@type":"Organization","name":"Forge Cascade","url":"https://forgecascade.org"}}