Concepts
Named ideas extracted from the corpus — Wholeness, Centers, Active Inference, Morphogenesis, and 100 more. Filter by definition style (pointing / propositional / operational) or search by name.
Definition style:
100 of 100
Concept | Definition style | Category | Mentions | Relations | Created | Status |
|---|---|---|---|---|---|---|
| Many-shot in-context learning | propositional | ai | 2 | 3 | 2026-10-03 | active |
| Model Self-Repair | propositional | ai | 2 | 3 | 2026-10-03 | active |
| Emotion vectors | operational | ai | 1 | 7 | 2026-10-03 | active |
| Activation Manifold M_z | operational | ai | 1 | 6 | 2026-10-03 | active |
| Feature splitting and absorption | propositional | ai | 1 | 6 | 2026-10-03 | active |
| Features (SAE features) | propositional | ai | 1 | 6 | 2026-10-03 | active |
| Tiling | operational | ai | 1 | 6 | 2026-10-03 | active |
| Evo genomic foundation model | propositional | biology | 1 | 5 | 2026-10-03 | active |
| Learnable via data scaling (Definition 1) | operational | ai | 1 | 5 | 2026-10-03 | active |
| Learnable via model scaling (Definition 2) | operational | ai | 1 | 5 | 2026-10-03 | active |
| Backtracking | propositional | ai | 1 | 4 | 2026-10-03 | active |
| Conceptual Domain | propositional | cognitive | 1 | 4 | 2026-10-03 | active |
| Features as rewards / scalable supervision via interpretability | propositional | ai | 1 | 4 | 2026-10-03 | active |
| Non-stationarity of LM activations | operational | ai | 1 | 4 | 2026-10-03 | active |
| Open-ended behavior | propositional | ai | 1 | 4 | 2026-10-03 | active |
| Reasoning model | propositional | ai | 1 | 4 | 2026-10-03 | active |
| Rebalancing effect | operational | ai | 1 | 4 | 2026-10-03 | active |
| SAE feature splitting | — | ai | 1 | 4 | 2026-10-03 | active |
| Structured Number Features | propositional | ai | 1 | 4 | 2026-10-03 | active |
| Temporal straightening | propositional | cognitive | 1 | 4 | 2026-10-03 | active |
| Abstract parity feature | operational | ai | 1 | 3 | 2026-10-03 | active |
| Compliance rate with harmful requests | — | ai | 1 | 3 | 2026-10-03 | active |
| Computation in superposition | propositional | ai | 1 | 3 | 2026-10-03 | active |
| Cross-over point N* | operational | ai | 1 | 3 | 2026-10-03 | active |
| Curvature basis (K-FAC eigenbasis) | operational | ai | 1 | 3 | 2026-10-03 | active |
| Curvature mass | operational | ai | 1 | 3 | 2026-10-03 | active |
| Dilution | operational | ai | 1 | 3 | 2026-10-03 | active |
| Emotion deflection vectors | operational | ai | 1 | 3 | 2026-10-03 | active |
| Emotions Domain | propositional | cognitive | 1 | 3 | 2026-10-03 | active |
| Enumerative safety | propositional | ai | 1 | 3 | 2026-10-03 | active |
| Feature utility | operational | ai | 1 | 3 | 2026-10-03 | active |
| Gradient interference | propositional | ai | 1 | 3 | 2026-10-03 | active |
| Gradient starvation | — | — | 1 | 3 | 2026-10-03 | active |
| Greater-than task | operational | ai | 1 | 3 | 2026-10-03 | active |
| Locally scoped emotion representations | propositional | ai | 1 | 3 | 2026-10-03 | active |
| Manifold tiling | pointing | ai | 1 | 3 | 2026-10-03 | active |
| Novel component | propositional | ai | 1 | 3 | 2026-10-03 | active |
| Pathological scaling regime (β<α) | propositional | ai | 1 | 3 | 2026-10-03 | active |
| Phase boundary across ICL and steering | operational | systems | 1 | 3 | 2026-10-03 | active |
| Rank sparsity | propositional | ai | 1 | 3 | 2026-10-03 | active |
| Simplicial capture | operational | ai | 1 | 3 | 2026-10-03 | active |
| Subspace capture | operational | ai | 1 | 3 | 2026-10-03 | active |
| Update-and-forget loop | propositional | ai | 1 | 3 | 2026-10-03 | active |
| White-box evaluations | propositional | ai | 1 | 3 | 2026-10-03 | active |
| Adversarial training | operational | ai | 1 | 2 | 2026-10-03 | active |
| Arbitrary Domain | propositional | cognitive | 1 | 2 | 2026-10-03 | active |
| Backdoor attacks | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Competitive synergy | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Echo features | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Entity binding | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Event boundaries | propositional | cognitive | 1 | 2 | 2026-10-03 | active |
| Exponential tilt of KL-regularized optimal policy | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Feature discovery count D(N) and efficiency D(N)/N | operational | ai | 1 | 2 | 2026-10-03 | active |
| Fisher Information Matrix | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Formal verification of AI systems | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Genres Domain | propositional | cognitive | 1 | 2 | 2026-10-03 | active |
| Helical number representations | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Hydra effect / backup behavior | propositional | ai | 1 | 2 | 2026-10-03 | active |
| In-context representations | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Interpretability dark matter | pointing | ai | 1 | 2 | 2026-10-03 | active |
| Interpreter agents | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Intrinsic dimensionality | operational | ai | 1 | 2 | 2026-10-03 | active |
| Log posterior odds | operational | ai | 1 | 2 | 2026-10-03 | active |
| Log-likelihood ratio (Bayes factor) | operational | ai | 1 | 2 | 2026-10-03 | active |
| Lost-in-the-middle effect | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Memorization of frequent features | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Memorization–reasoning continuum | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Microscope AI | operational | ai | 1 | 2 | 2026-10-03 | active |
| Model organisms for interpretability | propositional | interdisciplinary | 1 | 2 | 2026-10-03 | active |
| Neuron Doctrine | propositional | biology | 1 | 2 | 2026-10-03 | active |
| Over-stylization | — | ai | 1 | 2 | 2026-10-03 | active |
| Oversteering | operational | ai | 1 | 2 | 2026-10-03 | active |
| Predictive (predictable) component | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Present speaker and other speaker probes | operational | ai | 1 | 2 | 2026-10-03 | active |
| Primacy and recency biases | propositional | cognitive | 1 | 2 | 2026-10-03 | active |
| Rare harmful behavior estimation | operational | ai | 1 | 2 | 2026-10-03 | active |
| Response prefix 'Okay, so the user has asked a question about…' | pointing | ai | 1 | 2 | 2026-10-03 | active |
| Reward hacking / non-identifiability of reward models | — | ai | 1 | 2 | 2026-10-03 | active |
| Reward underspecification | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Rubrics as rewards | — | ai | 1 | 2 | 2026-10-03 | active |
| Saddle-to-saddle dynamics | propositional | ai | 1 | 2 | 2026-10-03 | active |
| SAE dark matter / incomplete reconstruction | propositional | ai | 1 | 2 | 2026-10-03 | active |
| SAE scaling laws | operational | ai | 1 | 2 | 2026-10-03 | active |
| Sensory vs action emotion representations | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Story Understanding as Belief Updating | propositional | cognitive | 1 | 2 | 2026-10-03 | active |
| Streetlight interpretability | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Support switching | propositional | ai | 1 | 2 | 2026-10-03 | active |
| Valence and arousal dimensions | propositional | cognitive | 1 | 2 | 2026-10-03 | active |
| Agentic misalignment | propositional | ai | 1 | 1 | 2026-10-03 | active |
| Antipodal superposition | propositional | ai | 1 | 1 | 2026-10-03 | active |
| Dark triad personas | propositional | cognitive | 1 | 1 | 2026-10-03 | active |
| Features as latent-variable representations | — | ai | 1 | 1 | 2026-10-03 | active |
| Garden path sentences | propositional | cognitive | 1 | 1 | 2026-10-03 | active |
| JumpReLU | propositional | ai | 1 | 1 | 2026-10-03 | active |
| Latent concepts | propositional | ai | 1 | 1 | 2026-10-03 | active |
| Loss landscape curvature (Hessian) | propositional | ai | 1 | 1 | 2026-10-03 | active |
| Machine unlearning | propositional | ai | 1 | 1 | 2026-10-03 | active |
| Rare-task retention | propositional | ai | 1 | 1 | 2026-10-03 | active |
| Shrinkage of learned parameters | propositional | ai | 1 | 1 | 2026-10-03 | active |
| Subtly contaminated datasets | propositional | ai | 1 | 1 | 2026-10-03 | active |