AI ethics evidence index Search
Source-linked daily record
Evidence record 7393 · automatically gathered

The Curse and Blessing of Mean Bias in FP4-Quantized LLM Training

FP4 training promises substantial memory and compute savings for large language models, but remains fragile because blockwise quantization is dictated by extreme activation magnitudes, which inflate dynamic range and compress long-tail signals. We identify a counterintuitive source of this failure: dominant activation outliers are not merely arbitrary sparse events, but are largely induced by a coherent rank-one mean bias, whose direction aligns with the leading anisotropic spectral component. T

Record details

Published: 11 March 2026
Source: arXiv
Category: Research
Topics: Bias & fairness
Retrieved: 14 July 2026

source-onlyevidence status

These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.

Shared topic: Bias & fairness

VulnAgent-R2: Evidence-Calibrated Multi-Agent Auditing for Repository-Level Vulnerability Detection

arXiv · 11 March 2026

Shared topic: Bias & fairness

On the Learning Dynamics of Two-layer Linear Networks with Label Noise SGD

arXiv · 11 March 2026

Shared topic: Bias & fairness

Mitigating Translationese Bias in Multilingual LLM-as-a-Judge via Disentangled Information Bottleneck

arXiv · 11 March 2026

Shared topic: Bias & fairness

Does Reasoning Make Search More Fair? Comparing Fairness in Reasoning and Non-Reasoning Rerankers

arXiv · 11 March 2026

Shared topic: Bias & fairness

The Algorithmic Blind Spot: Bias, Moral Status, and the Future of Robot Rights

arXiv · 10 March 2026

Shared topic: Bias & fairness

Procedural Fairness via Group Counterfactual Explanation

arXiv · 11 March 2026

How to cite this record

ethics.ai (11 March 2026), “The Curse and Blessing of Mean Bias in FP4-Quantized LLM Training,” evidence record 7393, https://ethics.ai/record/7393 (originally published by arXiv).

JSON

Use and limitations

This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.