AI300
  1. Mehtaab Sawhney / OpenAI for Science
    841.
    I've recently gone on leave from Columbia to join OpenAI, working on OpenAI for Science.›
    LLM score 75 · 2 months ago
  2. Ben Burtenshaw / Hugging Face Researcher
    842.
    We got Claude to teach open models how to write CUDA kernels.›
    LLM score 85 · 2 months ago
  3. alphaXiv
    843.
    BIG new idea in interpretability called Patterning›
    LLM score 75 · 2 months ago
  4. John Carmack
    844.
    #PaperADay 12 2019: Learning Latent Dynamics for Planning from Pixels (PlaNet)›
    LLM score 85 · 2 months ago
  5. Boris Cherny / Creator of Claude Code
    845.
    @charles_irl In case it’s not clear in the docs: - Ancestor https://t.co/v4FOLUBHz9’s are loaded into context automatically on startup›
    LLM score 20 · 2 months ago
  6. Lucas Beyer / Meta Researcher
    846.
    Another nice work with student-teacher setup, but this time with meta learning setup: teacher generates q/a pairs for student, and gets reward depending on student's improvement after training on those qa.›
    LLM score 75 · 2 months ago
  7. alphaXiv
    847.
    "LLM-in-Sandbox Elicits General Agentic Intelligence"›
    LLM score 85 · 2 months ago
  8. Ethan Shen / Ai2 Researcher
    848.
    @unsorsodicorda @saurabh_shah2 @Tim_Dettmers GA is trained using 4.5 Air as the teacher, whereas SERA-32B uses GLM 4.6.›
    LLM score 85 · 2 months ago
  9. Lucas Beyer / Meta Researcher
    849.
    I really like this one, it's a neat idea for self-improvement: distill token-level from the teacher is the student who was allowed to peek at the solution first.›
    LLM score 80 · 2 months ago
  10. Tim Dettmers / Research Scientist at Ai2
    850.
    his work was mostly the genius of Ethan Shen.›
    LLM score 70 · 2 months ago
  11. Tim Dettmers / Research Scientist at Ai2
    851.
    Our method became so efficient (26x vs RL; 57x vs other synth gen), that we could easily generate 1000s of trajectories for a single repo.›
    LLM score 85 · 2 months ago
  12. Tim Dettmers / Research Scientist at Ai2
    852.
    This is very impactful: you can now distill frontier performance into small models that are specialized to private repositories.›
    LLM score 75 · 2 months ago
  13. Tim Dettmers / Research Scientist at Ai2
    853.
    From there we could do a massiv amounts of experiments and really understand what matters for training coding agents.›
    LLM score 80 · 2 months ago
  14. Cameron Wolfe / Researcher at Netflix
    854.
    Continual learning is a popular topic in LLM research, but it might not be as far away as we think.›
    LLM score 65 · 2 months ago
  15. Ethan Shen / Ai2 Researcher
    855.
    Finally, we conduct an analysis of variance across SWE-Bench runs.›
    LLM score 95 · 2 months ago
  16. Ethan Shen / Ai2 Researcher
    856.
    We also experiment with mixing data from rollouts and discover that training on SERA’s first and second rollouts complement each other to increase performance in data constrained regimes.›
    LLM score 85 · 2 months ago
  17. Ethan Shen / Ai2 Researcher
    857.
    Even more interesting is the fact that the best data comes from truncated trajectories that just barely exceed the context limit.›
    LLM score 85 · 2 months ago
  18. Niklas Muennighoff / AI Researcher at Stanford
    858.
    Community-built open benchmarks work really well, e.g., Terminal-Bench, HLE, MMTEB.›
    LLM score 80 · 2 months ago
  19. John Carmack
    859.
    #PaperADay 11 Discovering state-of-the-art reinforcement learning algorithms›
    LLM score 80 · 2 months ago
  20. Noam Brown / OpenAI Research Scientist
    860.
    Had to cut this one for space: 2019: AI can't create art—creativity is uniquely human
    LLM score 20 · 2 months ago
  21. Andrej Karpathy / AI researcher
    861.
    @0xabi96 It feels like I’m cheating.›
    LLM score 70 · 2 months ago
  22. Andrej Karpathy / AI researcher
    862.
    @ChiragLathiya The nearest neighbor really is some kind of a junior engineer.›
    LLM score 80 · 2 months ago
  23. Andrej Karpathy / AI researcher
    863.
    @jeremytwei Love the word "comprehension debt", haven't encountered it so far, it's very accurate.›
    LLM score 30 · 2 months ago
  24. Andrej Karpathy / AI researcher
    864.
    @airesearch12 💯 @ Spec-driven development It's the limit of imperative -> declarative transition, basically being declarative entirely.›
    LLM score 60 · 2 months ago
  25. Andrej Karpathy / AI researcher
    865.
    A few random notes from claude coding quite a bit last few weeks.›
    LLM score 20 · 2 months ago
  26. Noam Brown / OpenAI Research Scientist
    866.
    1987: AI can't win at chess—planning is uniquely human›
    LLM score 70 · 2 months ago
  27. Yoshua Bengio
    867.
    Had a great discussion on AI's major societal impacts and emerging risks on the @RestIsPolitics podcast this past November.›
    LLM score 20 · 2 months ago
  28. Jonathan Ross / TPU Creator
    868.
    Success in the Information Age was about being able to answer questions.›
    LLM score 20 · 2 months ago
  29. Ben Burtenshaw / Hugging Face Researcher
    869.
    this is a blog post on claude + llama.cpp https://t.co/yej6WsNnQA
    LLM score 20 · 2 months ago
  30. Yoshua Bengio
    870.
    Merci à @HugoDecrypte pour l'invitation à faire un tour d'horizon des risques de l'IA ainsi que des solutions nécessaires pour se diriger vers une meilleure trajectoire.›
    LLM score 20 · 2 months ago
More