OthelloGPT https://www.lesswrong.com/posts/nmxzr2zsjNtjaHh7x/actually-othello-gpt-has-a-linear-emergent-world Weak to strong generalization https://youtu.be/OR-vcVNXdKk?si=aJIxY2LpcL_CX4de https://youtu.be/wBPZNhw1LV4?si=tZzqVnDFIn3cAmXe https://youtu.be/x_giZl0Zdx8?si=RBU1VKM0S9a4Ymsm Gpt4o Moshi https://x.com/JulianSlzr/status/1810303916686577858?t=0Rb6ITAlKWqJ1Zii1jqzpg&s=19 https://arxiv.org/abs/2407.06438 Important ML topics https://www.reddit.com/r/MachineLearning/s/vDV271v1yS Llms for math https://x.com/teortaxesTex/status/1810479422962597999?t=1yZ2Gyj_N-aps6QS9QDZQg&s=19 https://aimoprize.com/ https://fxtwitter.com/Thom_Wolf/status/1809895886899585164?t=57Zl4N1dg0MYFZbR2JDjyg&s=19 https://x.com/reach_vb/status/1811069858584363374?t=DcSeIzCCJtlLn-rcr6Ze5A&s=19 https://x.com/_lewtun/status/1811061090723201366 Deep learning leet code https://www.deep-ml.com/ Scientific machine learning https://www.reddit.com/r/MachineLearning/s/GozVlsoE1C Zeta alpha LLM papers https://www.youtube.com/watch?v=jCc3HaMcZ8Y https://www.youtube.com/@AICoffeeBreak Continual learning https://arxiv.org/abs/2310.01365 AlphaFold 3 https://x.com/ElanaPearl/status/1811053094664573039?t=A-95cnthVaYpXvLJ8Ihkgg&s=19 https://elanapearl.github.io/blog/2024/the-illustrated-alphafold/ https://www.youtube.com/@Eigensteve Steve Brunton The Math of Intelligence mostly machine learning https://www.youtube.com/playlist?list=PL2-dafEMk2A7mu0bSksCGMJEmeddU_H4D Data Driven Science & Engineering, Machine Learning, Dynamical Systems, and Control, Steve Brunton https://databookuw.com/databook.pdf Vision llms https://x.com/simonw/status/1811247901630505097?t=YtOcOZuLk_z26lqe9p5Haw&s=19 Survey of mixture of experts https://twitter.com/omarsar0/status/1811127876819026283?s=61&t=cUxlJsK7DrArhYjIs7W2_Q Transformer length generalization capabilities on many algorithmic tasks: addition, multiplication, in-context simulation of SGD https://arxiv.org/abs/2407.03310 Mathematical superintelligence is slowly but surely under construction https://x.com/burny_tech/status/1811538806539190624 https://x.com/Thom_Wolf/status/1809895886899585164 AI Mathematical Olympiad https://aimoprize.com/ https://x.com/reach_vb/status/1811069858584363374 https://x.com/teortaxesTex/status/1810881199025574077?t=48qM5aQTzyPNOg3j3aq6OQ&s=19 https://x.com/HarmonicMath/status/1810765353389281346 https://www.harmonic.fun/news https://deepmind.google/discover/blog/alphageometry-an-olympiad-level-ai-system-for-geometry/ https://www.nature.com/articles/s41586-023-06747-5 https://arxiv.org/abs/2404.06405 https://deepmind.google/discover/blog/funsearch-making-new-discoveries-in-mathematical-sciences-using-large-language-models/ https://www.nature.com/articles/s41586-023-06924-6 https://arxiv.org/abs/2006.08381 GraphRAG https://x.com/emileifrem/status/1811450877254758413?s=46&t=pNiPoM95FfYEFMSA5jsCMA Internet of agents https://arxiv.org/abs/2407.07061 Mixture of million experts https://arxiv.org/abs/2407.04153 If you're working on ARC-AGI, you should take a look at Sebastien Ferre's approach. It feels conceptually closer to human reasoning, compared to other program synthesis methods. Tackling the Abstraction and Reasoning Corpus (ARC) with Object-centric Models and the MDL Principle https://arxiv.org/abs/2311.00545 Intersection between artificial and biological intelligence https://www.cell.com/neuron/fulltext/S0896-6273(17)30509-3?_returnURL=https%3A%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS0896627317305093%3Fshowall%3Dtrue https://www.sciencedirect.com/science/article/pii/S0893608021003683?via%3Dihub https://arxiv.org/abs/1905.00547 https://portlandpress.com/essaysbiochem/article/60/4/381/78407/Synthetic-biology-routes-to-bio-artificial elastic weight consolidation (EWC) algorithm https://www.lesswrong.com/posts/Qn3ZDf9WAqGuAjWQe/on-scalable-oversight-with-weak-llms-judging-strong-llms Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems https://arxiv.org/abs/2405.06624 https://www.youtube.com/watch?v=mwUfAvzjd8o A New Physics-Inspired Theory of Deep Learning | Optimal initialization of Neural Nets https://www.youtube.com/watch?v=m2bXL5Z5CBM&t=778s&pp=ygUhcGh5c2ljc3QgaGVvcnkgb2YgbmV1cmFsIG5ldHdvcmtz Accelerated Training by Amplifying Slow Gradients https://www.youtube.com/watch?v=__xQw60y200 An Exactly Solvable Model for Emergence and Scaling Laws https://www.youtube.com/watch?v=-i8rqo32aIM https://arxiv.org/abs/2404.17563 Exploring Learning Dynamics in Concept Space https://www.youtube.com/watch?v=FoAGwdQUALM https://arxiv.org/abs/2406.19370v1 Memory^3: Language Modeling with Explicit Memory https://arxiv.org/abs/2407.01178 Mrchinterp circular features https://arxiv.org/html/2405.14860v1 anthropic models mechanistic interpretability fractal dimension https://youtu.be/Nlkk3glap_U?si=qb6KGDzK89KZT6Lt Training data for causal reasoning https://arxiv.org/abs/2407.07612 Nerf neural radience fields https://youtu.be/CRlN-cYFxTk?si=L5ZQjS7AtbRKjj84 Reverse engineering biological worm "C. elegans generates a well-defined repertoire of motor programs: locomotion, feeding, head oscillations, defecation, egg-laying, and postural changes." https://www.sciencedirect.com/science/article/pii/S0092867423008504#:~:text=The%20Caenorhabditis%20elegans%20nervous%20system%20consists%20of%20302%20neurons%20with%20known%20connectivity.&text=11%2C12%2C13-,C.,%2Dlaying%2C%20and%20postural%20changes https://x.com/ChombaBupe/status/1812140970646863895?t=N9gSQuXdhHa6d6in11iugw&s=19 Gauge theory https://youtu.be/_NY_Oi_BMIA?si=7JgoJ8BtkZ4q2PVg Brain flexibility https://www.sciencedirect.com/science/article/pii/S2352154624000020 ![[1720743568191.2001447528.png]]