Toppertube › Programming & Tech

Safety Alignment Should be Made More Than Just a Few Tokens Deep (Paper Explained)

Yannic Kilcher · Programming & Tech · 48 min · 14 thousand views · published 2024-12-10

This paper demonstrates in a series of experiments that current safety alignment techniques of LLMs, as well as corresponding jailbreaking a deep learning machine learning arxiv explained neural networks ai artificial in

Watch this lesson free on Toppertube, take notes beside the video, and add it to a course to track your progress and earn a certificate.

More from Yannic Kilcher

• I BUILT A FULLY AUTOMATIC MANSPLAINER Yannic Kilcher
• Traditional X-Mas Stream Yannic Kilcher
• TiDAR: Think in Diffusion, Talk in Autoregression (Paper Analysis) Yannic Kilcher
• Titans: Learning to Memorize at Test Time (Paper Analysis) Yannic Kilcher
• [Paper Analysis] The Free Transformer (and some Variational Autoencoder stuff) Yannic Kilcher
• [Video Response] What Cloudflare's code mode misses about MCP and tool calling Yannic Kilcher
• [Paper Analysis] On the Theoretical Limitations of Embedding-Based Retrieval (Warning: Rant) Yannic Kilcher
• AGI is not coming! Yannic Kilcher

More in Programming & Tech

• The Agents Were Fine. Everything Around Them Broke. - Joey Wang, Harvey AI Engineer
• 60% of Our Data PRs Now Merge With No Human Review - Chetas Joshi, Robinhood AI Engineer
• An Agent That Fixes Its Own Mistakes - Saloni Gandhi & Ziwei Guo, LinkedIn AI Engineer
• Stop Chatting With Spreadsheets — Francesco Galletta & Felippe Felisola Caso, Monzo AI Engineer
• Your Self-Hosted LLM Needs to Say No - Harshul Jain & Parv Khatri, Audible & SAP AI Engineer
• OpenAI spent $1.2 million hacking Linux... and it was a bargain Fireship
• How Gemini 4 Argon & Carbon were built (ft. DeepMind’s Head of Gemini Product) Google Cloud Tech
• How Are People Making Games Like This? Traversy Media