SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2604.07955 · 2026

Rethinking Residual Errors in Compensation-based LLM Quantization

Hong Gu, Haibin Shen, Kejie Huang, Juncan Deng, Shuaiting Li, Kedong Xu, Rongtao Deng, Minghan Jiang

arXiv · PDF · Open in the Atlas

Code that ran

3 of 8 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; 2 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 1 failure of our instrument; 5 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/2604.07955.json · badge: /api/ran/2604.07955.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
list0830/ResComp canonical (classifier label)
classifier basis: census
3 ran · 5 no recorded run
FunctionStatusWhere it lives
get_rope_function_name Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
list0830/ResComp/fake_quant/model_utils.py
pointer only (licence: NONE) · get_code("8005fc23ccbe61fb")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, list0830/ResComp 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.

llama_down_proj_groupsize Ran
ran, honoured the contract we drafted
list0830/ResComp/fake_quant/utils.py
pointer only (licence: NONE) · get_code("fe39dd44b7d2698b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, list0830/ResComp 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.

random_orthogonal_matrix Ran
ran, honoured the contract we drafted
list0830/ResComp/fake_quant/rotation_utils.py
pointer only (licence: NONE) · get_code("522f6e6438fa5941")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, list0830/ResComp 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.

cleanup_memory Not yet run
unverified (harvested, no recorded run)
list0830/ResComp/fake_quant/utils.py
pointer only (licence: NONE); the stored copy get_code serves, from intelligent-computing-lab-yale/gptaq, is pointer only (licence: NONE) · get_code("88c8d591b4688832")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, list0830/ResComp 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.

config_logging Not yet run
unverified (harvested, no recorded run)
list0830/ResComp/fake_quant/utils.py
pointer only (licence: NONE) · get_code("cf7f28fedcdd19c5")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, list0830/ResComp 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_layers Not yet run
unverified (harvested, no recorded run)
list0830/ResComp/fake_quant/model_utils.py
pointer only (licence: NONE) · get_code("eb59d0a73734e449")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, list0830/ResComp 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_wikitext2 Not yet run
unverified (harvested, no recorded run)
list0830/ResComp/fake_quant/data_utils.py
pointer only (licence: NONE) · get_code("347f7e0a0b0e4f2f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, list0830/ResComp 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.

model_type_extractor Not yet run
unverified (harvested, no recorded run)
list0830/ResComp/fake_quant/model_utils.py
pointer only (licence: NONE) · get_code("cdd805f7fc7a0842")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, list0830/ResComp 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 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 7 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 0 the same month, 7 later) and 1 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
get_rope_function_name 4 other papers
3 dated earlier, 1 later
2404.00456 (2024-04)
QuaRot: Outlier-Free 4-Bit Inference in Rotated LLMs
2511.04063 (2025-11)
DartQuant: Efficient Rotational Distribution Calibration for LLM Quantization
2602.02546 (2026-02)
D 2 Quant: Accurate Low-bit Post-Training Weight Quantization for LLMs
and 1 more not named here
cleanup_memory 3 other papers
2 dated earlier, 1 later
2504.02692 (2025-04)
GPTAQ: Efficient Finetuning-Free Quantization for Asymmetric Calibration
2505.13866 (2025-05)
Reasoning Path Compression: Compressing Generation Trajectories for Efficient LLM Reasoning
2605.15491 (2026-05)
Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs
get_layers 3 other papers
2 dated earlier, 1 later
2404.00456 (2024-04)
QuaRot: Outlier-Free 4-Bit Inference in Rotated LLMs
2511.04063 (2025-11)
DartQuant: Efficient Rotational Distribution Calibration for LLM Quantization
2605.15491 (2026-05)
Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs
model_type_extractor 3 other papers
2 dated earlier, 1 later
2404.00456 (2024-04)
QuaRot: Outlier-Free 4-Bit Inference in Rotated LLMs
2511.04063 (2025-11)
DartQuant: Efficient Rotational Distribution Calibration for LLM Quantization
2605.15491 (2026-05)
Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs
random_orthogonal_matrix 2 other papers
1 dated earlier, 1 later
2405.16406 (2024-05)
SpinQuant: LLM quantization with learned rotations
2605.15491 (2026-05)
Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs
config_logging 1 other paper
1 later
2605.15491 (2026-05)
Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs
llama_down_proj_groupsize 1 other paper
1 later
2605.15491 (2026-05)
Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs

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

Methods based on weight compensation, which iteratively apply quantization and weight compensation to minimize the output error, have recently demonstrated remarkable success in quantizing Large Language Models (LLMs). The representative work, GPTQ, introduces several key techniques that make such iterative methods practical for LLMs with billions of parameters. GPTAQ extends this approach by introducing an asymmetric calibration process that aligns the output of each quantized layer with its full-precision counterpart, incorporating a residual error into the weight compensation framework. In this work, we revisit the formulation of the residual error. We identify a sub-optimal calibration objective in existing methods: during the intra-layer calibration process, they align the quantized output with the output from compensated weights, rather than the true output from the original full-precision model. Therefore, we redefine the objective to precisely align the quantized model's output with the original output of the full-precision model at each step. We then reveal that the residual error originates not only from the output difference of the preceding layer but also from the discrepancy between the compensated and original weights within each layer, which we name the 'compensation-aware error'. By inheriting the neuron decomposition technique from GPTAQ, we can efficiently incorporate this compensation-aware error into the weight update process. Extensive experiments on various LLMs and quantization settings demonstrate that our proposed enhancements integrate seamlessly with both GPTQ and GPTAQ, significantly improving their quantization performance. Our code is publicly available at https://github.com/list0830/ResComp.

For agents

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

get_harvested_code_for_paper("2604.07955")
get_code_for_paper("2604.07955")
have("2604.07955")

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

curl https://syntology.ai/api/ran/2604.07955.json

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

[![Syntology run record](https://syntology.ai/api/ran/2604.07955.svg)](https://syntology.ai/paper/2604.07955)

Connect an agent — have() is free.