Papers › Resolving Discrepancies in Compute-Optimal Scaling of Language Models

Resolving Discrepancies in Compute-Optimal Scaling of Language Models

27 Jun 2024arXiv:2406.19146archive 2025-07-28

Tomer Porian, Mitchell Wortsman, Jenia Jitsev, Ludwig Schmidt, Yair Carmon

Kaplan et al. and Hoffmann et al. developed influential scaling laws for the optimal model size as a function of the compute budget, but these laws yield substantially different predictions. We explain the discrepancy by reproducing the Kaplan scaling law on two datasets (OpenWebText2 and RefinedWeb) and identifying three factors causing the difference: last layer computational cost, warmup duration, and scale-dependent optimizer tuning. With these factors corrected, we obtain excellent agreement with the Hoffmann et al. (i.e., "Chinchilla") scaling law. Counter to a hypothesis of Hoffmann et al., we find that careful learning rate decay is not essential for the validity of their scaling law. As a secondary result, we derive scaling laws for the optimal learning rate and batch size, finding that tuning the AdamW β₂ parameter is essential at lower batch sizes.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 16 code samples harvested from 3 repositories linked to this paper: 10 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 2 honoured, 3 violated, 0 with no contract checked; 5 where Syntology's instrument failed) · 6 unverified. Where Syntology's instrument failed: 4 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 15 samples from 2 repositories, 9 ran (4 with no instrument failure; 5 where Syntology's instrument failed); community (archive-listed): 1 sample from 1 repository, 1 ran (1 with no instrument failure; 0 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 2 of the 10 samples that ran have no recorded date.

formll/resolving-scaling-law-discrepancies officialmentioned in papermentioned on GitHubSyntology: 3 samples, 3 ran (1 instrument-failed) · 0 unverified report
formll/resolving-scaling-law-discrepencies officialmentioned in papermentioned on GitHubMITSyntology: 12 samples, 6 ran (4 instrument-failed) · 6 unverified report
hadasah/scaling_laws mentioned on GitHubpytorchSyntology: 1 sample, 1 ran (0 instrument-failed) · 0 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

16 samples harvested: 10 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 2 honoured, 3 violated, 0 with no contract checked; 5 where Syntology's instrument failed) · 6 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

2ran · honoured contract
3ran · violated contract
4ran · our draft was wrong
1ran; the contract check could not run
6unverified

Licence: 4 of the 16 samples are pointer only, meaning Syntology does not serve that copy's text. For 3 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 3 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. 1 sample whose link to this paper records no repository is listed under one of this paper's own repositories, because identical code was harvested from it for another paper. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

formll/resolving-scaling-law-discrepancies official repository 3 samples, 3 ran (1 instrument-failed) · 0 unverified
custom_huber_loss analysis.py ran · violated contract fingerprinted MIT (permissive) · 752782550c94c617 · report fixture: signature y_true:ndarray/1/float, y_pred:ndarray/1/float, delta:float/0/float · returned float64 · level recorded 2026-09-22
fetch_flop analysis.py ran · honoured contract MIT (permissive) · 15bb4fddf111e4b7 · report fixture: signature df:DataFrame/0/?, flop:float/0/float, loss_key:str/0/str, warmup_remove_factor:float/0/float, n_key:str/0/str, seq_len:int/0/int, bs_key:str/0/str, keep_bs_lr_keys:bool/0/bool, flop_per_token_key:str/0/str, flop_tolerance:float/0/float · returned DataFrame · level recorded 2026-09-22
huber_loss_objective analysis.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant fingerprinted MIT (permissive) · cd1ae572668e4ce4 · report fixture: signature params:list/1/float, F:ndarray/1/float, losses:ndarray/1/float · returned float64 · level recorded 2026-09-22
formll/resolving-scaling-law-discrepencies official repository 12 samples, 6 ran (4 instrument-failed) · 6 unverified
add_zero paper_tables.py ran · our draft was wrong fingerprinted MIT (permissive) · 6a004a1784aa8d01 · report fixture: signature x:str/0/str · returned str · level recorded 2026-09-24
get_color in 28 papers utils.py ran · our draft was wrong fingerprinted MIT (permissive) here; the stored copy is pointer only · 8d1a5df6af7bf845 · report fixture: signature i:int/0/int, clist:list/1/str · returned str · level recorded, date not recorded
get_linestyle utils.py ran · our draft was wrong fingerprinted MIT (permissive) · 99378ea8da320872 · report fixture: signature i:int/0/int · returned str · level recorded 2026-09-24
get_marker utils.py ran · our draft was wrong fingerprinted MIT (permissive) · 9bcb697d9afad429 · report fixture: signature i:int/0/int · returned str · level recorded 2026-09-24
precise_flops_per_token_chinchilla data.py ran · violated contract fingerprinted MIT (permissive) here; the stored copy is pointer only · fc9bc7439b7fd07b · report fixture: signature width:ndarray/1/int, depth:ndarray/1/int · returned ndarray · level recorded 2026-09-22
precise_param_count_open_lm data.py ran · honoured contract MIT (permissive) here; the stored copy is pointer only · e20ba47c892c9025 · report fixture: signature width:int64/0/int, depth:int64/0/int, vocab_size:int/0/int · returned int64 · level recorded 2026-09-22
draw_fit plotting.py unverified MIT (permissive) · 4fa5926c92763b45 · report
extract_CI paper_tables.py unverified MIT (permissive) · 4c878b8671db62ea · report
extract_label in 12 papers paper_tables.py unverified MIT (permissive) · 9370894d3db38de0 · report
isoflop_curves_plot plotting.py unverified MIT (permissive) · 51aa561ef23bceaa · report
opt_param_vs_compute_plot plotting.py unverified MIT (permissive) · 3ae2bda49e8d6539 · report
proportional_sliding_window_filter data.py unverified MIT (permissive) · b8859ce442ad4d05 · report
hadasah/scaling_laws community (archive-listed) 1 sample, 1 ran (0 instrument-failed) · 0 unverified
maybe_get_item paper_analysis_and_plots.py identical code Syntology harvested from this repository for another paper; this paper's link records no repository, path or licence of its own ran · violated contract fingerprinted licence of this copy not recorded · a6d9e09c75556988 · report fixture: signature x:Tensor/2/float, i:int/0/int · returned Tensor · level recorded, date not recorded

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

AdamW

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections