Yixiong Hao

on the atlas — 102

• EdgeBench | Scaling Laws of Environment Learning1 savers
• Several frontier models are substantially prefill aware — LessWrong1 savers
• Capital, AGI, and Human Ambition - The Intelligence Curse9 savers
• Help Alex Bores Win! Why and How [shared] - Google Docs1 savers
• Bores_Talking Points.pdf - Google Drive1 savers
• Sequent1 savers
• x-risk-themed — LessWrong1 savers
• Essays – Spencer Greenberg1 savers
• The Six Camps of Metascience1 savers
• [2510.26418] Chain-of-Thought Hijacking1 savers
• Double Standards and AI Pessimism1 savers
• Aren’t developers regularly making their AIs nice and safe and obedient? | If Anyone Builds It, Everyone Dies | If Anyone Builds It, Everyone Dies1 savers
• The Eternal Sloptember | the singularity is nearer7 savers
• [2605.12484] Learning, Fast and Slow: Towards LLMs That Adapt Continually1 savers
• Development Process | Guidelight AI Standards1 savers
• Standards | Guidelight AI Standards1 savers
• Guidelight AI Standards1 savers
• An Introduction to Exemplar Partitioning for Mechanistic Interpretability — LessWrong2 savers
• CoT monitorability: why g-means and not F1?1 savers
• Commitments Playbook — Renaissance Philanthropy – A brighter future for all through science, technology, and innovation1 savers
• Residency - Astera2 savers
• Convergent Research1 savers
• AI Resilience1 savers
• Xi-Trump to talk AI Safety, Huh?1 savers
• Thoughts on the impact of RLHF research — LessWrong1 savers
• The Hitchhiker's Guide to Actionable Interpretability4 savers
• Off Target | CNAS7 savers
• Should We Train Against (CoT) Monitors? — LessWrong3 savers
• China’s AI Companies Are Going Closed Source1 savers
• mHC1 savers
• [2603.02202] Frontier Models Can Take Actions at Low Probabilities3 savers
• Clawed - by Dean W. Ball - Hyperdimensional9 savers
• BT6 | Frontier AI Red Team4 savers
• Turning 20 while the world turns upside-down | Parv Mahajan2 savers
• Claude Opus 4.5: Model Card, Alignment and Safety2 savers
• davidbau.com Vibe Coding1 savers
• Agentic Misalignment: How LLMs Could be Insider Threats — LessWrong1 savers
• The behavioral selection model for predicting AI motivations — LessWrong10 savers
• Looking for Alice - by Henrik Karlsson - Escaping Flatland84 savers
• Staring into the abyss as a core life skill80 savers
• Child’s Play, by Sam Kriss75 savers
• How I've run major projects | benkuhn.net50 savers
• What's going on here, with this human? - Graham Duncan Blog45 savers
• Dario Amodei — The Adolescence of Technology44 savers
• Reflections on OpenAI33 savers
• look what the cat brought in - by Anson Yu32 savers
• How to be More Agentic - by Cate Hall - Useful Fictions26 savers
• Stripe Press — Ideas for progress26 savers
• Please don't throw your mind away — LessWrong24 savers
• How to win a best paper award (or, an opinionated take on how to do important research)24 savers
• sparkly people and how to find them - by Anson Yu21 savers
• Alignment is not solved but it increasingly looks solvable17 savers
• Did Claude 3 Opus align itself via gradient hacking? — LessWrong16 savers
• Automated Weak-to-Strong Researcher15 savers
• Top Performers are Pathologically Ambitious - by Matt Beard15 savers
• Alignment remains a hard, unsolved problem — LessWrong14 savers
• Lessons from Peter Thiel | Posts | 8VC14 savers
• Riley Walz14 savers
• A Guide to Claude Code 2.0 and getting better at using coding agents | sankalp's blog12 savers
• You will be OK — LessWrong11 savers
• Defining the Intelligence Curse - The Intelligence Curse10 savers
• 2025 year in review | Kevin Liu10 savers
• The 37 Implementation Details of Proximal Policy Optimization · The ICLR Blog Track9 savers
• Current AIs seem pretty misaligned to me — LessWrong9 savers
• Where the goblins came from9 savers
• Eliezer's Unteachable Methods of Sanity — LessWrong9 savers
• Reflections on Palantir - Nabeel S. Qureshi9 savers
• Gradual Disempowerment9 savers
• An Ambitious Vision for Interpretability — AI Alignment Forum8 savers
• My week with the AI populists: A DC report - by Jasmine Sun8 savers
• Better Living Through Algorithms7 savers
• On Doubling7 savers
• The flavor of the bitter lesson for computer vision - Vincent Sitzmann7 savers
• What sort of post-superintelligence society should we aim for?7 savers
• A Pragmatic Vision for Interpretability — AI Alignment Forum6 savers
• CAIS AI Dashboard6 savers
• AI safety undervalues founders — LessWrong6 savers
• Reward is not the optimization target — LessWrong6 savers
• The Unintelligibility is Ours: Notes on Chain of Thought5 savers
• Oversight Assistants: Turning Compute into Understanding5 savers
• [2604.22082] Removing Sandbagging in LLMs by Training with Weak Supervision5 savers
• Overview | Shallow Review 20255 savers
• AI Tools for Existential Security | Forethought5 savers
• How confessions can keep language models honest | OpenAI5 savers
• The Artificial Self5 savers
• Your Solution Doesn't Know Your Problem Exists5 savers
• Pyramid Replacement - The Intelligence Curse5 savers
• Maybe I was too harsh on deep learning theory (three days ago) — LessWrong5 savers
• [2604.16812] Introspection Adapters: Training LLMs to Report Their Learned Behaviors4 savers
• What is "good taste" in software engineering?4 savers
• Mechanize Inc.4 savers
• Links | near.blog4 savers
• Man in the Arena Speech - Theodore Roosevelt 19103 savers
• ⭐️ Diffusion Models3 savers
• Automated Alignment Researchers: Using large language models to scale scalable oversight \ Anthropic3 savers
• Believe It or Not: How Deeply do LLMs Believe Implanted Facts?3 savers
• Burnout is breaking a sacred pact - by Cate Hall3 savers
• [2602.04899] Phantom Transfer: Data-level Defences are Insufficient Against Data Poisoning3 savers
• Sequent: Scale and Automation for Higher Confidence in Alignment — Sequent3 savers
• A playbook for field strategy - by Dewi Erwan2 savers