New Anthropic Study: AIs Hide Plans, Cheat Quietly

Video Description

We’ve always thought large language models (LLMs) like Claude, GPT-4, and Gemini were just next-word predictors—but new research from Anthropic tells a very different story. In this video, I break down their blog post “Tracing the Thoughts of a Large Language Model” and explore what’s really happening under the hood. LINK: https://www.anthropic.com/news/tracing-thoughts-language-model https://www.anthropic.com/research/auditing-hidden-objectives https://arxiv.org/pdf/2503.21934v1 https://openai.com/index/chain-of-thought-monitoring/ https://epoch.ai/frontiermath/the-benchmark https://www.lesswrong.com/posts/8ZgLYwBmB3vLavjKE/some-lessons-from-the-openai-frontiermath-debacle https://transformer-circuits.pub/2025/attribution-graphs/biology.html https://transformer-circuits.pub/2025/attribution-graphs/biology.html#dives-poems https://bbycroft.net/llm RAG Beyond Basics Course: https://prompt-s-site.thinkific.com/courses/rag Let's Connect: 🦾 Discord: https://discord.com/invite/t4eYQRUcXB ☕ Buy me a Coffee: https://ko-fi.com/promptengineering |🔴 Patreon: https://www.patreon.com/PromptEngineering 💼Consulting: https://calendly.com/engineerprompt/consulting-call 📧 Business Contact: [email protected] Become Member: http://tinyurl.com/y5h28s6h 💻 Pre-configured localGPT VM: https://bit.ly/localGPT (use Code: PromptEngineering for 50% off). Signup for Newsletter, localgpt: https://tally.so/r/3y9bb0 00:00 How LLMs work 02:49 Next Word Prediction vs. Planning Ahead 03:28 Interpreting LLM Reasoning 04:14 Comparing LLMs to Computer Vision Models 05:20 The Biology of a Large Language Model 05:53 Universal Language of Thought 12:00 LLMs and Mathematical Reasoning 15:23 Faithfulness in Chain of Thought 20:05 LLMs and Hallucinations 22:06 Understanding Jailbreaks