concept
active
concept:learnable-via-model-scaling-definition-2Learnable via model scaling (Definition 2)
Loss a smaller model never reaches even asymptotically.
Neighborhood — ranked by edge-count
Papers (1)
paper
Concepts (1)
concept
- Learnable via data scaling (Definition 1)associated_withrelated_toLoss gains a smaller model recovers with infinite data.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Core goal statement.
- A representation that captures relevant aspects of a system; according to the theorem, the regulator must embody this.
- Opening question.
- Interpretive process for transforming many-valued contexts into formal contexts via scale attributes.
- Clarifies what unsupervised learning does.
- Authors argue features are model properties because logit effects and ablations are consistent with feature interpretations
- Used in the color cooccurrence experiment to embed colors into 3D space preserving dissimilarity matrix distances
- More capable models benefit more from Verbalized Sampling, showing an emergent scaling trendclaim0.728Empirical observation that larger models (GPT-4.1, Gemini-2.5-Pro) show 1.5-2x greater diversity gains from VS compared to smaller models