Skip to content

michalstrnadel/source-to-skill / examples/neural-networks-3blue1brown

neural-networks-3blue1brown

Verified

Course skill distilled from 3Blue1Brown's "Neural networks" YouTube playlist (10 videos) - from a plain MNIST digit classifier (neurons, weights, biases, sigmoid/ReLU), through cost functions, gradient descent, stochastic gradient descent and backpropagation with its chain-rule calculus, to large language models and transformers (tokens, embeddings, softmax and temperature, attention heads with query/key/value, masking, MLP blocks storing facts, superposition, the full GPT-3 175B parameter tally), cross-entropy loss, KL divergence and distillation, and diffusion image/video models (CLIP, DDPM, DDIM, classifier-free guidance). Load when explaining or teaching how neural networks, backprop, transformers, attention, LLM training losses or diffusion models work, when someone needs an intuition-first mental model or a concrete number (e.g. GPT-3 dimensions), or asks "what does this 3Blue1Brown lesson say about X".

evals/last-run.json

View on GitHub
{
  "date": "2026-10-26",
  "score": 1.0,
  "passed": 13,
  "total": 13,
  "grader": "blind: one subagent answered from the skill directory only; graded against evals.json drafted from re-extracted full_text.txt",
  "results": [
    {
      "id": 1,
      "passed": [
        true,
        true,
        true
      ],
      "evidence": [
        "\"784-16-16-10: 784 input neurons (28x28 pixels), two hidden layers of 16 neurons\" (lessons/01)",
        "\"about 13,000 weights and biases in total\" (lessons/01 &t=743s)",
        "\"about 96% right, and about 98% after a few tweaks\" (lessons/02 &t=750s, &t=816s)"
      ]
    },
    {
      "id": 2,
      "passed": [
        true,
        true
      ],
      "evidence": [
        "\"A human reading GPT-3's training text non-stop would need over 2,600 years\" (lessons/05)",
        "\"would take well over 100 million years\" (lessons/05 &t=187s)"
      ]
    },
    {
      "id": 3,
      "passed": [
        true,
        true
      ],
      "evidence": [
        "\"Summing every training example for every step is very slow ... gives a big computational speedup\" (lessons/03 &t=554s)",
        "\"each step only approximates the true gradient\" / \"stochastic gradient descent\" (&t=617s)"
      ]
    },
    {
      "id": 4,
      "passed": [
        true,
        true
      ],
      "evidence": [
        "\"Higher T flattens the distribution ... At T = 0 the model always picks the top token\" (lessons/06 &t=1417s)",
        "\"the API limits T to 2. This is an arbitrary product constraint\" (&t=1483s)"
      ]
    },
    {
      "id": 5,
      "passed": [
        true,
        true
      ],
      "evidence": [
        "\"96 layers give just under 58B, about 1/3 of the total\" (lessons/07 &t=1430s)",
        "\"96 blocks give ~116B, about 2/3 of the total\" (lessons/08 &t=984s)"
      ]
    },
    {
      "id": 6,
      "passed": [
        true,
        true
      ],
      "evidence": [
        "\"introduced in Lesson 8 ... at 17:27 (1047s)\" - 10 s from 1057 s",
        "\"If 'nearly perpendicular' (89°–91°) is allowed, high dimensions hold far more\" / grows exponentially with dimension"
      ]
    }
  ]
}
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexUnverifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall