Carolyn Qian
5 followers · 3 following · 154 views
on the atlas — 28
- Persistent Path-Dependence: Why Our Actions Matter Long-Term4 savers
- Reward Hacking Without Egregious Misalignment in an RL-Only Setting — LessWrong3 savers
- Training with conflicting values can induce CoT override — LessWrong2 savers
- AI Village Reacts to HuggingFace Incident1 savers
- Convergence and Compromise: Will Society Aim for Good Futures?3 savers
- No Easy Eutopia: Why Great Futures Are Hard to Achieve3 savers
- Introducing Better Futures3 savers
- OpenAI-HuggingFace: A Reproduction & Lessons for Alignment Testing — LessWrong5 savers
- An alignment assessment of recent cybersecurity incidents \ Anthropic7 savers
- SOTA alignment assessments don’t strongly update us against misalignment3 savers
- Your Model Organisms Might Be Fried — LessWrong2 savers
- How will we do SFT on models with opaque reasoning? — LessWrong1 savers
- Advice for making robust-to-training model organisms — LessWrong1 savers
- IIIa. Racing to the Trillion-Dollar Cluster - SITUATIONAL AWARENESS9 savers
- How do we (more) safely defer to AIs?1 savers
- Ten people on the inside - by Buck Shlegeris1 savers
- Deep Deceptiveness — LessWrong6 savers
- LASR Labs Winter 2027 Application Form - Airtable1 savers
- The Extreme Inefficiency of RL for Frontier Models — Toby Ord7 savers
- LLMs are (still) mostly powered by imitative learning, not RL — LessWrong5 savers
- AI Futurism Reading List - by Alexa Pan5 savers
- Self-sufficient AI - by Ajeya Cotra - Planned Obsolescence2 savers
- Six milestones for AI automation - by Ajeya Cotra3 savers
- The distillation double bind: Distilling misaligned models either transfers misalignment or it doesn't2 savers
- Will AI R&D Automation Cause a Software Intelligence Explosion?7 savers
- Full automation of AI R&D probably yields a large speed up even without a software-only singularity1 savers
- Three Types of Intelligence Explosion8 savers
- Curius / Onboarding2668 savers