RonaldTooTall, Anthropic warns: AI poisoning makes open models vulnerable "sleeper agents" that generate harmful code under specific triggers.
https://arstechnica.com/information-technology/2024/01/ai-poisoning-could-turn-open-models-into-destructive-sleeper-agents-says-anthropic/
#LLMs #AI #ArtificialIntelligence #LargeLanguageModels #CyberSecurity #Security #Technology