Subjective Multi-Bias Detection with Large Language Models
2026-08-10 • Computation and Language
Computation and Language
AI summaryⓘ
The authors worked on identifying different kinds of subjective bias in text, which can make statements seem unfair or untrue. They focused on three types: framing bias (showing one-sided views), epistemological bias (using language that affects how believable the text is), and demographic bias (language assumptions about groups like gender or religion). Using a dataset called WIKIBIAS with Wikipedia edits labeled for these biases, they built models to detect when these biases appear in text. Their project and code are openly available online.
subjective biasframing biasepistemological biasdemographic biastext classificationbias detectionWikipedia editsWIKIBIAS datasetmulti-span bias
Authors
Ruiyu Li, Zhiying Zhu
Abstract
In this project, we delved into the pervasive challenge of bias detection within the text content. More specifically, our focus lies on the identification of subjective bias, a type of bias that introduces improper attitudes or portrays a statement at odds with the actual truth. The subjective bias can jeopardize the authenticity and reliability of texts, leading to misconceptions and potential social tensions, especially when expressed through offensive language. Following prior work [1], we tackled with three different types of subjective biases in text: (1) framing bias with the use of one-sided words or phrases containing a particular point of view; (2) epistemological bias which includes subtle linguistic features that can affect the believability of the texts; (3) demographic bias with word/phrase usage under presuppositions of a particular demographic factor (i.e., gender or religion). In terms of the data we utilize, the input consists of texts that may harbor subjective biases. The output is a classification or annotation that reveals the presence or absence of such biases within the provided content. More specifically, we detected three different types of multi-span biases in corpus WIKIBIAS [2] with more than 4,000 sentence pairs from Wikipedia edits. The data is labelled by bias type for span pairs with the following categories: (1) framing bias, (2) epistemological bias, (3) demographic bias, and (4) no bias. The project codes are released at https://github.com/HoningJade/LLM-Bias-Type-Classification.