LLM Exchange Rates Updated - by Arctotherium

How do LLM's trade off lives between different categories?

Warning: article is almost entirely images. Update: Part 2 here. Update: Part 3 here. On February 19th, 2025, the Center for AI Safety published “Utility Engineering: Analyzing and Controlling Emergent Value Systems in AIs” (website, code, paper). In this paper, they showed that modern LLMs have coherent and transitive implicit utility functions and world models, and provided methods and code to extract them. Among other things, they showed that bigger and more capable LLMs had more coherent and more transitive (ie, preferring A > B and B > C implies A > C) preferences. Figure 16, which…

saved by

• Vincent Cheng

related reading

• LLM Exchange Rates Updated: #5arctotherium.substack.com
• gpt-4.pdfcdn.openai.com
• Noam Brown on X: "Implications of Large-Scale Test-Time Compute" / Xx.com
• What We’ve Learned From A Year of Building with LLMs – Applied LLMsapplied-llms.org
• The Big LLM Architecture Comparisonmagazine.sebastianraschka.com
• Things we learned about LLMs in 2024simonwillison.net
• Escalation Risks from Language Models in Military and Diplomatic Decision-Makingarxiv.org
• Paper AI Tigersgleech.org
• What language do language models speak?tcz.hu
• [2604.21751] Why are all LLMs Obsessed with Japanese Culture? On the Hidden Cultural and Regional Biases of LLMsarxiv.org
• Taking LLMs Seriously (As Language Models) — LessWronglesswrong.com
• [2607.14345] Value Leakage: An LLM's Answers Are Silently Shaped by Its Own Valuesarxiv.org