Pith. sign in

REVIEW 3 cited by

A Hitchhiker's Guide to Scaling Law Estimation

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2410.11840 v2 pith:VP5ADBCF submitted 2024-10-15 cs.LG cs.AIcs.CL

A Hitchhiker's Guide to Scaling Law Estimation

classification cs.LG cs.AIcs.CL
keywords modelscalinglawstrainingmodelsfamiliesbehaviorbest
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Scaling laws predict the loss of a target machine learning model by extrapolating from easier-to-train models with fewer parameters or smaller training sets. This provides an efficient way for practitioners and researchers alike to compare pretraining decisions involving optimizers, datasets, and model architectures. Despite the widespread use of scaling laws to model the dynamics of language model training, there has been little work on understanding how to best estimate and interpret them. We collect (and release) a large-scale dataset containing losses and downstream evaluations for 485 previously published pretrained models. We use these to estimate more than 1000 scaling laws, then derive a set of best practices for estimating scaling laws in new model families. We find that fitting scaling laws to intermediate checkpoints of training runs (and not just their final losses) substantially improves accuracy, and that -- all else equal -- estimates of performance are generally most accurate when derived from other models of similar sizes. However, because there is a significant degree of variability across model seeds, training multiple small models is sometimes more useful than training a single large one. Moreover, while different model families differ scaling behavior, they are often similar enough that a target model's behavior can be predicted from a single model with the same architecture, along with scaling parameter estimates derived from other model families.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. On the Invariance and Generality of Neural Scaling Laws

    cs.LG 2026-05 unverdicted novelty 7.0

    Neural scaling laws are invariant under bijective data transformations and change predictably with information resolution ρ under non-bijective transformations, enabling cross-domain transport of fitted exponents.

  2. Validity Threats for Foundation Model Research

    cs.LG 2026-06 accept novelty 6.0

    Maps common low-compute research strategies for foundation models onto statistical, internal, external, and construct validity threats via a causal-inference lens.

  3. Child-directed speech facilitates production, not comprehension, in BabyLMs

    cs.CL 2026-05 unverdicted novelty 6.0

    CDS-trained BabyLMs show earlier and more appropriate production in a new frame-completion task while FineWeb-edu models lead on comprehension benchmarks, indicating current tests underestimate CDS benefits.