Interpretability jargon, sorted into grug, gorm, and gormug tiers
Sauers_ · x · 2026-11-04
A playful taxonomy of interpretability research jargon in three tiers:
- grug style (plain): probes, J-lens, neuron atlas
- gorm style (fancy): activation oracles, introspection, LoRAcles
- gormug combo: SAEs, weight-sparse training, transcoders, parameter decomposition, BSF
A meme-y but informative tour of interpretability terminology.
More from Fun
- AI video imagines an Englishman in New York: tea, umbrella and a very confused vibe — MandelDuck · 2026-11-04
- Training an LLM Explained as a Cricket Match: AI Community Meme — cneuralnetwork · 2026-11-04
- Dev nostalgia: when our biggest problem was the Next.js dev server eating RAM — jonathan_wilke · 2026-11-04
- Chicago Kare: faithful reproduction of Susan Kare's 1984 Mac system font — genmon · 2026-11-04
- Vibe-coded with Opus 5.5 and Meshy: five stained-glass 3D scenes with switchable lighting — Pristine_Good7326 · 2026-11-04
- OpenAI DevDay demo shows forking agent task history at any step — paw_lean · 2026-11-04