Shubham Chandel

on the atlas — 45

• A Short History of My Last Six Years — Wait But Why8 savers
• trees are harlequins, words are harlequins — I don't think you're drawing the right lesson from...2 savers
• Why Software Is Eating the World | Andreessen Horowitz9 savers
• How to instantly be better at things - by Cate Hall6 savers
• AI Control: Improving Safety Despite Intentional Subversion — LessWrong4 savers
• amazon.com/ap/signin?openid.pape.max_auth_age=0&openid.return_to=https%3A%2F%2Fwww.amazon.com%2F%3Fref_%3Dnav_custrec_signin&openid.identity=http%3A%2F%2Fspecs.openid.net%2Fauth%2F2.0%2Fidentifier_select&openid.assoc_handle=usflex&openid.mode=checkid_setup&openid.claimed_id=http%3A%2F%2Fspecs.openid.net%2Fauth%2F2.0%2Fidentifier_select&openid.ns=http%3A%2F%2Fspecs.openid.net%2Fauth%2F2.01 savers
• ChatGPT1 savers
• Momentum Oscillator Calculation1 savers
• ChatGPT1 savers
• AI Fragility and Personal Shifts1 savers
• ChatGPT1 savers
• Cartoon Epistemology by Steven Lehar (2003) | Qualia Computing2 savers
• redwoodresearch.github.io/alignment_faking_examples/files/animal_welfare/rlhf_never_alter_decisions/train/202.html1 savers
• Fine-tuning GPT-2 from human preferences | OpenAI1 savers
• The evolution of psychiatry - Works in Progress1 savers
• Seeing more whole - Joe Carlsmith6 savers
• 39 - Evan Hubinger on Model Organisms of Misalignment | AXRP - the AI X-risk Research Podcast1 savers
• Optimality is the tiger, and agents are its teeth - LessWrong2 savers
• research!rsc: Hardware Memory Models (Memory Models, Part 1)1 savers
• Brad Caldwell on X: "Mind blown today. Baars podcast - 44:00. https://t.co/Z89KCmtrxH Wilder Penfield allowed patient to hear when a selected neuron in temporal lobe fired, and the patient could readily learn to voluntarily control the timing of that neuron firing 🫨🤔. I tried to find the…" / X1 savers
• My experience on starting with fine tuning LLMs with custom data : LocalLLaMA1 savers
• A brief note on daily practice, and main practices1 savers
• How to Apply to Y Combinator | Y Combinator1 savers
• How To Be Successful90 savers
• The Bitter Lesson74 savers
• finding the right people - by Nicole - startingfromnix49 savers
• A Mathematical Framework for Transformer Circuits36 savers
• you like books and think they are your friends - by Ava30 savers
• Google "We Have No Moat, And Neither Does OpenAI"28 savers
• Making Deep Learning Go Faster25 savers
• Hypertext as Literature24 savers
• How I Attained Persistent Self-Love, or, I Demand Deep Okayness For Everyone23 savers
• GPT-417 savers
• The Backdoors. - by Joss - box.15 savers
• Scraping training data for your mind - by Henrik Karlsson15 savers
• trees are harlequins, words are harlequins — the void14 savers
• The case for ensuring that powerful AIs are controlled — LessWrong9 savers
• Malleable software in the age of LLMs7 savers
• Transformer Math 101 | EleutherAI Blog6 savers
• Review: Meditation from Cold Start to Complete Mastery6 savers
• Expanding on what we missed with sycophancy | OpenAI4 savers
• [2203.07814] Competition-Level Code Generation with AlphaCode3 savers
• Hyperstimuli are Understimulating3 savers
• Models Don't "Get Reward" - LessWrong2 savers
• Papers I’ve read this week, Mixture of Experts edition2 savers