finding
active
finding:olmo-models-trained-with-larger-injection-gaps-n-1-10-20-50-100-reach-higher-task-loss-despite-equal-global-frequencyOLMo models trained with larger injection gaps N∈{1,10,20,50,100} reach higher task loss despite equal global frequency
Fig. 6b.
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
- Discussion and App. A.
Hypotheses (1)
hypothesis
- Hypothesis box in §3.2.
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.
- Larger OLMo models achieve both low training loss and high test accuracy, i.e. learn generalizable structurefinding0.853Fig. 5.
- Fig. 5.
- Fig. 22.
- Fig. 4.
- Data-induced competition over resources.
- Selective pressure toward convergence via task generality
- Core framing.
- Fig. 23.