New ASCII art attacks can jailbreak LLMs
New ASCII art attacks can jailbreak LLMs. Anthropic releases new AI models, Elon Musk sues OpenAI, and Google is taking steps to tackle AI spam content.
What a week of events in the AI space, let's dive into them.
đ° News:
âšī¸ ArtPrompt: ASCII Art-based Jailbreak Attacks against Aligned LLMs
âšī¸ Anthropic presents the Claude 3 family. Three new large language models with impressive benchmarks.
âšī¸ Here Comes The AI Worm: Unleashing Zero-click Worms
âšī¸ New ways of tackling spammy, low-quality content on Search | Google
âšī¸ Elon Musk sues OpenAI for abandoning original mission for profit
âšī¸ Adobe previews new generative AI tools for crafting and editing custom audio