IIIS AI Safety Lunch
A monthly lunch on AI safety at the FIT Building, Tsinghua University, Beijing.
If you are based in Beijing, work on AI safety, and would like to join, please email Tianxing He with a brief introduction and a CV. You are welcome to write in Chinese.
2026
When Generative Models Become Game Engines: A Brief Introduction to Genie
Interactive Game Video Generation: An Introduction to Hunyuan-GameCraft
Yi Xu
Reverse-Engineering Model Editing on Language Models
Explainability and Trace Detection in LLM Unlearning
Yixuan Wu
On Safety Issues of Large Language Model Multi-Agent Systems
Ziheng Zhou
Can Large Language Models Reinvent Foundational Algorithms?
2025
Digital Protection in the Era of Generative Models
Unveiling the Basin-Like Loss Landscape in Large Language Models
When Bias Pretends to Be Truth: How Spurious Correlations Undermine Hallucination Detection in LLMs
AI Safety in Context
Pay for the Second-Best Service: A Game-Theoretic Approach Against Dishonest LLM Providers
Interpretability Analysis of LLMs with a Mixture-of-Experts Architecture
Liuxian Ma
Analyzing Catastrophic Risks and Cognitive Abilities in LM Agents
Several Simple Attacks and Findings from Black-Box Interactions with GPT-OSS
Chengda Lu