finding
active
finding:larger-olmo-models-achieve-both-low-training-loss-and-high-test-accuracy-i-e-learn-generalizable-structureLarger OLMo models achieve both low training loss and high test accuracy, i.e. learn generalizable structure
Fig. 5.
Source paper
extracted_from(2026) · Jing Huang · Daniel Wurgaft · Rachit Bansal · Laura Ruis +5
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Data-induced competition over resources.
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.
- Fig. 6b.
- Hypothesis box in §3.2.
- Fig. 5.
- Fig. 22.
- Core framing.
- Selective pressure toward convergence via task generality
- Are there model structures that account for memorization and generalization across large swaths of training data?question0.777Aggregate analysis motivation.
- Training on image data should improve LLM performance, and training on language data should improve vision model performancehypothesis0.775Implication of PRH for cross-modal training efficiency