SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2602.02522 · 2026

IMU-1: Sample-Efficient Pre-training of Small Language Models

George Grigorev

arXiv · PDF · Open in the Atlas

Code that ran

6 of 10 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 6 failures of our instrument; 4 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2602.02522.json · badge: /api/ran/2602.02522.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
thepowerfuldeez/sample_efficient_gpt canonical (classifier label)
classifier basis: census
6 ran · 4 no recorded run
FunctionStatusWhere it lives
first_pad Ran
ran; our contract draft was wrong, not the code
thepowerfuldeez/sample_efficient_gpt/sample_efficient_gpt/evals/chat_web.py
pointer only (licence: NONE) · get_code("48944aa712166327")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, thepowerfuldeez/sample_efficient_gpt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

render_prompts_lm Ran
ran; our contract draft was wrong, not the code
thepowerfuldeez/sample_efficient_gpt/sample_efficient_gpt/evals/core_eval.py
pointer only (licence: NONE); the stored copy get_code serves, from serdardoesml/recursive-lm, is pointer only (licence: NONE) · get_code("6519cb0a9ba0f137")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, thepowerfuldeez/sample_efficient_gpt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

render_prompts_mc Ran
ran; our contract draft was wrong, not the code
thepowerfuldeez/sample_efficient_gpt/sample_efficient_gpt/evals/core_eval.py
pointer only (licence: NONE); the stored copy get_code serves, from serdardoesml/recursive-lm, is pointer only (licence: NONE) · get_code("038adb078ed905c9")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, thepowerfuldeez/sample_efficient_gpt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

render_prompts_schema Ran
ran; our contract draft was wrong, not the code
thepowerfuldeez/sample_efficient_gpt/sample_efficient_gpt/evals/core_eval.py
pointer only (licence: NONE); the stored copy get_code serves, from serdardoesml/recursive-lm, is pointer only (licence: NONE) · get_code("d72ac88289f6e9aa")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, thepowerfuldeez/sample_efficient_gpt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

softmax Ran
ran; our contract draft was wrong, not the code
thepowerfuldeez/sample_efficient_gpt/sample_efficient_gpt/transformer/core.py
pointer only (licence: NONE) · get_code("fb874de8e38ead7a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, thepowerfuldeez/sample_efficient_gpt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

tokenize Ran
ran; our contract draft was wrong, not the code
thepowerfuldeez/sample_efficient_gpt/sample_efficient_gpt/evals/chat_web.py
pointer only (licence: NONE) · get_code("4e22c143139d965f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, thepowerfuldeez/sample_efficient_gpt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

execute_code Not yet run
unverified (harvested, no recorded run)
thepowerfuldeez/sample_efficient_gpt/sample_efficient_gpt/evals/execution.py
pointer only (licence: NONE) · get_code("686a4f93a88bb3d2")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, thepowerfuldeez/sample_efficient_gpt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

get_model Not yet run
unverified (harvested, no recorded run)
thepowerfuldeez/sample_efficient_gpt/sample_efficient_gpt/evals/chat_web.py
pointer only (licence: NONE) · get_code("f2a76e19103b5cb1")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, thepowerfuldeez/sample_efficient_gpt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

get_model_hf Not yet run
unverified (harvested, no recorded run)
thepowerfuldeez/sample_efficient_gpt/sample_efficient_gpt/evals/base_eval.py
pointer only (licence: NONE) · get_code("ae3a6a9b1cb51538")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, thepowerfuldeez/sample_efficient_gpt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

load_chat_template Not yet run
unverified (harvested, no recorded run)
thepowerfuldeez/sample_efficient_gpt/sample_efficient_gpt/evals/chat_eval.py
pointer only (licence: NONE) · get_code("d5243a7a6c949a5e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, thepowerfuldeez/sample_efficient_gpt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 4 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 4 later) and 6 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
render_prompts_lm 3 other papers
3 later
2605.17842 (2026-05)
SNLP: Layer-Parallel Inference via Structured Newton Corrections
2606.00761 (2026-06)
Confidence-Adaptive SwiGLU for Mixture-of-Experts
2608.26973 (2026-08)
Squeezing More from Limited Data with Recursive Transformers
render_prompts_mc 3 other papers
3 later
2605.17842 (2026-05)
SNLP: Layer-Parallel Inference via Structured Newton Corrections
2606.00761 (2026-06)
Confidence-Adaptive SwiGLU for Mixture-of-Experts
2608.26973 (2026-08)
Squeezing More from Limited Data with Recursive Transformers
render_prompts_schema 3 other papers
3 later
2605.17842 (2026-05)
SNLP: Layer-Parallel Inference via Structured Newton Corrections
2606.00761 (2026-06)
Confidence-Adaptive SwiGLU for Mixture-of-Experts
2608.26973 (2026-08)
Squeezing More from Limited Data with Recursive Transformers
execute_code 1 other paper
1 later
2605.17842 (2026-05)
SNLP: Layer-Parallel Inference via Structured Newton Corrections

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We present IMU-1, a 430M-parameter language model trained on 72B tokens that approaches the benchmark performance of models trained on 56× more data. We describe a validated training recipe combining recent architectural interventions (QK-norm attention, per-head gating, value residuals, LayerNorm scaling) with optimization advances (NorMuon with cautious weight decay, µP parametrization) and a three-stage training schedule with post-hoc checkpoint EMA. We provide ablations for each component and release code, weights and data to enable reproduction: https://huggingface.co/thepowerfuldeez/imu1_base

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2602.02522")
get_code_for_paper("2602.02522")
have("2602.02522")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2602.02522.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2602.02522.svg)](https://syntology.ai/paper/2602.02522)

Connect an agent — have() is free.