Papers › Same Geometry, Opposite Noise: Transformer Magnitude Representations Lack Scalar Variability

Same Geometry, Opposite Noise: Transformer Magnitude Representations Lack Scalar Variability

6 Apr 2026arXiv:2604.04469added by Syntology

Jon-Paul Cacioli

Title, abstract, authors and date from arXiv's metadata (CC0); this paper is not in the Papers with Code archive (frozen 2025-07-28).

Scalar variability -- the finding that representational noise scales proportionally with magnitude, producing a constant coefficient of variation -- is a hallmark of biological magnitude systems. We tested whether transformer language models exhibit this property by analysing the dispersion of hidden-state representations across carrier sentences for 26 numerical magnitudes in three 7-8B parameter models (Llama-3-8B-Instruct, Mistral-7B-Instruct-v0.3, Llama-3-8B-Base; data from Cacioli, 2026). We found the opposite: representational variability decreased with magnitude along the magnitude axis (scaling exponent alpha approx -0.19; 0/16 primary layers with alpha > 0, all three models). The negative sign was consistent in full-dimensional space (alpha approx -0.04) and after sentence-identity correction (alpha approx -0.007). The anti-scalar pattern was 3-5x stronger along the magnitude axis than orthogonal dimensions, and corpus frequency strongly predicted per-magnitude variability (rho = .84). These results demonstrate that distributional learning alone is insufficient to produce scalar variability: transformers reproduce log-compressive magnitude geometry but not the constant-CV noise signature observed in biological systems.

PaperPDFCode 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 25 code samples harvested from 1 repository linked to this paper: 17 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 2 honoured, 1 violated, 4 with no contract checked; 10 where Syntology's instrument failed) · 8 unverified. Where Syntology's instrument failed: 8 where our contract draft was wrong; 2 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: found in paper text by Syntology: 25 samples from 1 repository, 17 ran (7 with no instrument failure; 10 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 on 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

synthiumjp/weber found in paper text by SyntologySyntology: 25 samples, 17 ran (10 instrument-failed) · 8 unverified report

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

25 samples harvested: 17 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 2 honoured, 1 violated, 4 with no contract checked; 10 where Syntology's instrument failed) · 8 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
1ran · violated contract
8ran · our draft was wrong
2ran; the contract check could not run
4ran
8unverified

Licence: 25 of the 25 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from synthiumjp/weber. “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.

synthiumjp/weber found in paper text by Syntology 25 samples, 17 ran (10 instrument-failed) · 8 unverified
bootstrap_dprime_difference m3_pilot/m3_discrimination_analysis.py ran no licence file found · pointer only · 0164081b18d1d18f · report fixture: signature cross_trials:list/1/dict, within_trials:list/1/dict, n_boot:int/0/int, seed:int/0/int · returned dict · level recorded 2026-09-25
build_chat_prompt m3_pilot/m3_discrimination.py ran no licence file found · pointer only · 86978420de191da1 · report fixture: signature tokenizer:MockTokenizer/0/?, question:str/0/str, use_chat_template:bool/0/bool · returned str · level recorded 2026-09-25
build_prompt in 78 papers m3_pilot/m3_rerun_identification.py ran · our draft was wrong fingerprinted no licence file found · pointer only · a192415d50000694 · report fixture: signature question:str/0/str, option_a:str/0/str, option_b:str/0/str · returned str · level recorded 2026-09-25
build_trial_prompts m3_pilot/m3_discrimination_stimuli.py ran · our draft was wrong no licence file found · pointer only · 2945b23ed816b1b1 · report fixture: signature trial:dict/0/? · returned dict · level recorded 2026-09-25
check_existing_outputs m3_pilot/m3_batch_run.py ran · violated contract fingerprinted no licence file found · pointer only · ba591e2b8e0a3b68 · report fixture: signature model_short:str/0/str, condition:str/0/str · returned bool · level recorded 2026-09-25
compute_empirical_rdm m3_pilot/m3_pilot_analysis.py ran; the contract check could not run (RAISES) no licence file found · pointer only · 9f1b6862d6dd1240 · report fixture: signature centroids:ndarray/2/float, metric:str/0/str · returned ndarray · level recorded 2026-09-25
dprime_2afc m3_pilot/m3_discrimination_analysis.py ran · honoured contract fingerprinted no licence file found · pointer only · 0b1c72331ceec341 · report fixture: signature p_correct:float/0/float · returned float64 · level recorded 2026-09-25
dprime_from_trials m3_pilot/m3_discrimination_analysis.py ran · honoured contract no licence file found · pointer only · dfbff91f5a4e7f21 · report fixture: signature trials:list/1/dict · returned float64 · level recorded 2026-09-25
format_duration in 11 papers m3_pilot/m3_batch_run.py ran · our draft was wrong fingerprinted no licence file found · pointer only · 47e2bf1a6a4e395a · report fixture: signature seconds:float/0/float · returned str · level recorded 2026-09-25
format_value_for_prompt m3_pilot/m3_discrimination_stimuli.py ran · our draft was wrong fingerprinted no licence file found · pointer only · 9d9cc36448d8441f · report fixture: signature x:float/0/float · returned str · level recorded 2026-09-25
generate_base_pairs m3_pilot/m3_discrimination_stimuli.py ran · our draft was wrong no licence file found · pointer only · 98de5458330aab2a · report fixture: signature position:str/0/str, target_log_dist:float/0/float, rng:Generator/0/? · returned list · level recorded 2026-09-25
get_ab_token_ids m3_pilot/m3_discrimination.py ran · our draft was wrong no licence file found · pointer only · 943fdac52a0174b7 · report fixture: signature tokenizer:MinimalTokenizer/0/? · returned tuple · level recorded 2026-09-25
load_centroids m3_pilot/m3_pilot_analysis.py ran · our draft was wrong no licence file found · pointer only · 42f29f084b292a99 · report fixture: signature config:AnalysisConfig/0/?, condition:str/0/str · returned dict · level recorded 2026-09-25
load_stimuli m3_pilot/m3_pilot_analysis.py ran · our draft was wrong no licence file found · pointer only · 82dc0bac01224528 · report fixture: signature config:AnalysisConfig/0/?, condition:str/0/str · returned dict · level recorded 2026-09-25
run_extraction m3_pilot/m3_batch_run.py ran no licence file found · pointer only · fec0e0aa2fd81e02 · report fixture: signature model:dict/0/?, condition:str/0/str, dry_run:bool/0/bool, skip_identification:bool/0/bool · returned tuple · level recorded 2026-09-25
train_category_probe m3_pilot/m3_causal.py ran; the contract check could not run (RAISES) no licence file found · pointer only · fdd634fece88536f · report fixture: signature centroids:ndarray/3/float, values:ndarray/1/float, boundary:int/0/int · returned dict · level recorded 2026-09-25
validate_direction_pca m3_pilot/m3_causal.py ran no licence file found · pointer only · 39a694b8fae5fbf6 · report fixture: signature centroids:ndarray/3/float, values:ndarray/1/float, directions:ndarray/2/float, boundary:int/0/int · returned dict · level recorded 2026-09-25
extract_ab_logits m3_pilot/m3_rerun_identification.py unverified no licence file found · pointer only · 447464de65e29478 · report
find_magnitude_token_position m3_pilot/m3_extract.py unverified no licence file found · pointer only · 076c40365db781e5 · report
load_centroids_for_probe m3_pilot/m3_causal.py unverified no licence file found · pointer only · a767a15c18a627d3 · report
load_data in 700 papers m3_pilot/m3_exploratory.py unverified no licence file found · pointer only · eb7d32c97737c623 · report
load_model_and_tokenizer in 61 papers m3_pilot/m3_extract.py unverified no licence file found · pointer only · 6e731249ef09bbf4 · report
run_counterbalanced_trial m3_pilot/m3_rerun_identification.py unverified no licence file found · pointer only · 448a5903c062d1a0 · report
run_single_trial m3_pilot/m3_discrimination.py unverified no licence file found · pointer only · 91265a7aecc5e015 · report
verify_all_positions m3_pilot/m3_extract.py unverified no licence file found · pointer only · e9f6455c3534c26d · report

Results from the paper

The Papers with Code archive ends with its 2025-07-28 snapshot. This paper's arXiv identifier, 2604.04469, was issued in April 2026, after that date, so the archive has no leaderboard rows for it.

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