A two-step parametric model trained on small 'ladder' language models predicts individual task accuracy of 7B and 13B models to within 2 points on four of eight benchmarks, using about 1% of target compute.
Olmo 2: The best fully open language model to date, 2024
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Establishing Task Scaling Laws via Compute-Efficient Model Ladders
A two-step parametric model trained on small 'ladder' language models predicts individual task accuracy of 7B and 13B models to within 2 points on four of eight benchmarks, using about 1% of target compute.