On the Relationship between Truth and Political Bias in Language Models

• URL: http://arxiv.org/abs/2409.05283v2
• Date: Fri, 11 Oct 2024 20:10:53 GMT
• Title: On the Relationship between Truth and Political Bias in Language Models
• Authors: Suyash Fulay, William Brannon, Shrestha Mohanty, Cassandra Overney, Elinor Poole-Dayan, Deb Roy, Jad Kabbara,
• Abstract summary: We focus on analyzing the relationship between two concepts essential in both language model alignment and political science. We train reward models on various popular truthfulness datasets and evaluate their political bias. Our findings reveal that optimizing reward models for truthfulness on these datasets tends to result in a left-leaning political bias.
• Score: 22.57096615768638
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Language model alignment research often attempts to ensure that models are not only helpful and harmless, but also truthful and unbiased. However, optimizing these objectives simultaneously can obscure how improving one aspect might impact the others. In this work, we focus on analyzing the relationship between two concepts essential in both language model alignment and political science: truthfulness and political bias. We train reward models on various popular truthfulness datasets and subsequently evaluate their political bias. Our findings reveal that optimizing reward models for truthfulness on these datasets tends to result in a left-leaning political bias. We also find that existing open-source reward models (i.e., those trained on standard human preference datasets) already show a similar bias and that the bias is larger for larger models. These results raise important questions about the datasets used to represent truthfulness, potential limitations of aligning models to be both truthful and politically unbiased, and what language models capture about the relationship between truth and politics.

Related papers

• Balancing Transparency and Accuracy: A Comparative Analysis of Rule-Based and Deep Learning Models in Political Bias Classification [5.550237524713089]
  The study highlights the sensitivity of modern self-learning systems to unconstrained data ingestion. Applying both models to left-leaning (CNN) and right-leaning (FOX) news articles, we assess their effectiveness on data beyond the original training and test sets. We contrast the opaque architecture of a deep learning model with the transparency of a linguistically informed rule-based model.
  arXiv  Detail & Related papers  (2024-11-07T00:09:18Z)
• Spoken Stereoset: On Evaluating Social Bias Toward Speaker in Speech Large Language Models [50.40276881893513]
  This study introduces Spoken Stereoset, a dataset specifically designed to evaluate social biases in Speech Large Language Models (SLLMs) By examining how different models respond to speech from diverse demographic groups, we aim to identify these biases. The findings indicate that while most models show minimal bias, some still exhibit slightly stereotypical or anti-stereotypical tendencies.
  arXiv  Detail & Related papers  (2024-08-14T16:55:06Z)
• Representation Bias in Political Sample Simulations with Large Language Models [54.48283690603358]
  This study seeks to identify and quantify biases in simulating political samples with Large Language Models. Using the GPT-3.5-Turbo model, we leverage data from the American National Election Studies, German Longitudinal Election Study, Zuobiao dataset, and China Family Panel Studies.
  arXiv  Detail & Related papers  (2024-07-16T05:52:26Z)
• Fast Model Debias with Machine Unlearning [54.32026474971696]
  Deep neural networks might behave in a biased manner in many real-world scenarios. Existing debiasing methods suffer from high costs in bias labeling or model re-training. We propose a fast model debiasing framework (FMD) which offers an efficient approach to identify, evaluate and remove biases.
  arXiv  Detail & Related papers  (2023-10-19T08:10:57Z)
• It's All Relative: Interpretable Models for Scoring Bias in Documents [10.678219157857946]
  We propose an interpretable model to score the bias present in web documents, based only on their textual content. Our model incorporates assumptions reminiscent of the Bradley-Terry axioms and is trained on pairs of revisions of the same Wikipedia article. We show that we can interpret the parameters of the trained model to discover the words most indicative of bias.
  arXiv  Detail & Related papers  (2023-07-16T19:35:38Z)
• Debiasing Vision-Language Models via Biased Prompts [79.04467131711775]
  We propose a general approach for debiasing vision-language foundation models by projecting out biased directions in the text embedding. We show that debiasing only the text embedding with a calibrated projection matrix suffices to yield robust classifiers and fair generative models.
  arXiv  Detail & Related papers  (2023-01-31T20:09:33Z)
• RedditBias: A Real-World Resource for Bias Evaluation and Debiasing of Conversational Language Models [37.98671828283487]
  Text representation models are prone to exhibit a range of societal biases. Recent work has predominantly focused on measuring and mitigating bias in pretrained language models. We present RedditBias, the first conversational data set grounded in the actual human conversations from Reddit.
  arXiv  Detail & Related papers  (2021-06-07T11:22:39Z)
• Mitigating Political Bias in Language Models Through Reinforced Calibration [6.964628305312507]
  We describe metrics for measuring political bias in GPT-2 generation. We propose a reinforcement learning (RL) framework for mitigating political biases in generated text.
  arXiv  Detail & Related papers  (2021-04-30T07:21:30Z)
• Learning from others' mistakes: Avoiding dataset biases without modeling them [111.17078939377313]
  State-of-the-art natural language processing (NLP) models often learn to model dataset biases and surface form correlations instead of features that target the intended task. Previous work has demonstrated effective methods to circumvent these issues when knowledge of the bias is available. We show a method for training models that learn to ignore these problematic correlations.
  arXiv  Detail & Related papers  (2020-12-02T16:10:54Z)
• Inflating Topic Relevance with Ideology: A Case Study of Political Ideology Bias in Social Topic Detection Models [16.279854003220418]
  We investigate the impact of political ideology biases in training data. Our work highlights the susceptibility of large, complex models to propagating the biases from human-selected input. As a way to mitigate the bias, we propose to learn a text representation that is invariant to political ideology while still judging topic relevance.
  arXiv  Detail & Related papers  (2020-11-29T05:54:03Z)
• Towards Robustifying NLI Models Against Lexical Dataset Biases [94.79704960296108]
  This paper explores both data-level and model-level debiasing methods to robustify models against lexical dataset biases. First, we debias the dataset through data augmentation and enhancement, but show that the model bias cannot be fully removed via this method. The second approach employs a bag-of-words sub-model to capture the features that are likely to exploit the bias and prevents the original model from learning these biased features.
  arXiv  Detail & Related papers  (2020-05-10T17:56:10Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.