Paper Reading Group | Linear Representation Hypothesis in the context of LLM Safety

​❗❗❗ Event location changed from ETH Main building to ETH ISTP UNO B 11, Universitätstrasse 41 — see instructions below.

💬 The Zurich AI Safety (ZAIS) group meets regularly to discuss topics in AI alignment and AI safety, with a general focus on recent research and the technical aspects and implications of such topics.

​🎓 This session: a special guest facilitation by Cynthia Chen, who will present her research: "Linear Representation Hypothesis in the context of LLM Safety."

​Cynthia is a PhD student at ETH Zurich, jointly supervised by Prof. Andreas Krause and Prof. Florian Tramèr. Before ETH, she studied at the University of Hong Kong and spent research time at UC Berkeley's Center for Human-Compatible AI, Stanford, and Columbia. Her current research focuses on the science of LLM alignment, learning human values, and reward hacking. More about her work: https://www.xccyn.com/

​🫵 Are you a student interested in technical alignment? This is a particularly great event for you. Cynthia has supervised semester projects and thesis work at ETH, so this is an excellent opportunity to meet her, hear about her research directions, and explore whether there might be a fit for your own studies.

​Who should come
As we often have lively discussions that rely on knowledge of key terms in the field, we recommend that you have some experience in ML, as well as some background in AI safety, for an engaging experience. If you have completed Zurich's AISF Programme or an analogous one, you are well-suited to join. A good way to check if you meet these criteria is to go through the syllabus of the AI Alignment course and see if you are familiar with these topics.


📍 ETH ISTP UNO B 11, Universitätstrasse 41 — People wait close to the door next to the glowing sign that reads "Einstein School" (location pin on Google) — we will let you in. Call the phone number in the Luma event instructions if you have a problem.