Do LLMs Have Beliefs and Desires? Exploring Cognition and Alignment
brwilder · x · 2026-08-05
The author explores whether LLMs possess cognitive states like beliefs and desires. Leaning towards 'yes', they argue that investigating this question reveals fascinating insights into how models solve cognitive tasks and what approaches to AI safety and alignment could actually work.
Related event: Exploring LLM Psychology: Do Models Have Beliefs and Preferences?(2 posts)→
More from AGI Musings
- US AI Firms Push to Slow Progress Just as Chinese Open-Source Catches Up — kevinnbass · 2026-08-05
- Introducing Pax Machina: A New Publication on Institutions for the AI Era — ryan_t_lowe · 2026-08-05
- Satirizing AI Alignment: A Viral Copypasta Mocking Super Alignment Jargon — voooooogel · 2026-08-05
- OpenAI Exec Joanne Jang: True Automation Requires Zero Thinking — joannejang · 2026-08-05
- Ashton Kutcher: AI Makes Dopamine Infinite, Physical Presence Becomes the Ultimate Scarcity — r0ck3t23 · 2026-08-05
- Building Software on the Eve of ASI: Six Core Opportunities — pzakin · 2026-08-05