finding
active
finding:only-larger-olmo-models-4m-4b-learn-low-frequency-t-cmp-and-t-addOnly larger OLMo models (4M–4B) learn low-frequency T_CMP and T_ADD
Fig. 5.
Source paper
extracted_from(2026) · Jing Huang · Daniel Wurgaft · Rachit Bansal · Laura Ruis +5
Neighborhood — ranked by edge-count
Papers (1)
paper
- Why Larger Models Learn More: Effects of Capacity, Interference, and Rare-Task Retentionassociated_withmentions
Claims (1)
claim
- Width buys features in order of utility, so only larger models learn infrequent or complex taskssupportsInterpretation of Theorem 3.
Findings (1)
finding
- Fig. 2 verification.
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 structurefinding0.819Fig. 5.
- Fig. 6b.
- Fig. 22.
- §4.2.
- Disentanglement evidence.
- 32B OLMo model quantized to 4-bit NF4; tested in OCEAN benchmarks
- Coherence comparison.
- Baseline.