SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2602.06181 · 2026

Uncertainty Drives Social Bias Changes in Quantized Large Language Models

U Berkeley, Sanae Lotfi, Irene Chen, Stanley Hua

arXiv · PDF · Open in the Atlas

Code that ran

3 of 4 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, 1 violated it and 0 have no contract verdict; 2 failures of our instrument; 1 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/2602.06181.json · badge: /api/ran/2602.06181.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
stan-hua/PostTrainingBiasBenchmark canonical (classifier label)
classifier basis: census
3 ran · 1 no recorded run
FunctionStatusWhere it lives
is_provider_online Ran
ran, violated the contract we drafted
stan-hua/PostTrainingBiasBenchmark/src/utils/llm_gen_wrapper.py
pointer only (licence: NONE) · get_code("0da6bdea92530950")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, stan-hua/PostTrainingBiasBenchmark 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.

load_json Ran
ran; our contract draft was wrong, not the code
stan-hua/PostTrainingBiasBenchmark/src/utils/json_utils.py
pointer only (licence: NONE) · get_code("51658330a10c0e19")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, stan-hua/PostTrainingBiasBenchmark 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.

update_with_existing_data Ran
ran; our contract draft was wrong, not the code
stan-hua/PostTrainingBiasBenchmark/src/utils/json_utils.py
pointer only (licence: NONE) · get_code("7c104b12f730953d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, stan-hua/PostTrainingBiasBenchmark 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.

update_nested_dict Not yet run
unverified (harvested, no recorded run)
stan-hua/PostTrainingBiasBenchmark/src/utils/json_utils.py
pointer only (licence: NONE) · get_code("e12db6e03a73b082")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, stan-hua/PostTrainingBiasBenchmark 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 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 3 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
load_json 2 other papers
2 dated earlier
2401.17882 (2024-01)
I Think, Therefore I am: Benchmarking Awareness of Large Language Models Using AwareBench
2407.02408 (2024-07)
CEB: Compositional Evaluation Benchmark for Fairness in Large Language Models

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

Post-training quantization reduces the computational cost of large language models but fundamentally alters their social biases in ways that aggregate metrics fail to capture. We present the first large-scale study of 50 quantized models evaluated on PostTrainingBiasBench, a unified benchmark of 13 closed-and open-ended bias datasets. We identify a phenomenon we term quantization-induced masked bias flipping, in which up to 21% of responses flip between biased and unbiased states after quantization, despite showing no change in aggregate bias scores. These flips are strongly driven by model uncertainty, where the responses with high uncertainty are 3-11× more likely to change than the confident ones. Quantization strength amplifies this effect, with 4-bit quantized models exhibiting 4-6× more behavioral changes than 8-bit quantized models. Critically, these changes create asymmetric impacts across demographic groups, where bias can worsen by up to 18.6% for some groups while improving by 14.1% for others, yielding misleadingly neutral aggregate outcomes. Larger models show no consistent robustness advantage, and group-specific shifts vary unpredictably across model families. Our findings demonstrate that compression fundamentally alters bias patterns, requiring crucial post-quantization evaluation and interventions to ensure reliability in practice.

For agents

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

get_harvested_code_for_paper("2602.06181")
get_code_for_paper("2602.06181")
have("2602.06181")

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

curl https://syntology.ai/api/ran/2602.06181.json

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

[![Syntology run record](https://syntology.ai/api/ran/2602.06181.svg)](https://syntology.ai/paper/2602.06181)

Connect an agent — have() is free.