Evaluating the Effects of Prompt Perturbation on Bias and Hallucination in Large Language Models
Summary
This arXiv paper evaluates how prompt perturbations affect bias and hallucination in LLMs used for decision-making tasks, finding that perturbations can mitigate these issues in some models, with Claude 3 outperforming GPT-3.5 across most datasets.
View Cached Full Text
Cached at: 09/30/26, 09:47 AM
# Evaluating the Effects of Prompt Perturbation on Bias and Hallucination in Large Language Models Source: [https://arxiv.org/abs/2609.35804](https://arxiv.org/abs/2609.35804) [View PDF](https://arxiv.org/pdf/2609.35804) > Abstract:Large language models \(LLMs\) have shown remarkable capabilities in various natural language processing tasks, leading to their widespread deployment as intelligent assistants in decision\-making contexts\. However, the increasing complexity of these models raises concerns about their reliability, particularly regarding bias and hallucination\. In this work, we evaluate the robustness of LLMs to perturbed variations of the original inquiry in decision\-making tasks\. We show that contrary to previous studies, perturbations can mitigate bias and hallucination in some LLMs over other models\. It's found that Claude 3 is more effective for the tasks represented in most datasets, whereas models like GPT3\.5 exhibit varying levels of adequacy, performing comparably in some cases but falling significantly behind in others\. These insights are crucial for understanding the practical implications of deploying LLM\-based assistants as effective decision\-support tools in real\-world applications, emphasising the need for rigorous testing and validation to ensure reliability and effectiveness\. This study contributes to the growing body of research on LLM evaluation and provides insights for developing more robust and trustworthy AI assistants in critical decision\-making contexts\. ## Submission history From: Ahmad Shahi \[[view email](https://arxiv.org/show-email/d048bff9/2609.35804)\] **\[v1\]**Fri, 18 Sep 2026 10:25:05 UTC \(119 KB\)
Similar Articles
Fluent but Fallible: Benchmarking Hallucination and Factual Reliability in Large Language Models
A comparative benchmark study evaluating ten LLMs on 70 academic prompts across four complexity levels, measuring hallucination and response deviation. Findings highlight a trade-off between response completeness and factual reliability, with Perplexity and DeepSeek V3 scoring highest on accuracy.
Mechanisms of Prompt-Induced Hallucination in Vision-Language Models
This paper investigates prompt-induced hallucinations in vision-language models through mechanistic analysis, identifying specific attention heads responsible for the models' tendency to favor textual prompts over visual evidence. The authors demonstrate that ablating these PIH-heads reduces hallucinations by at least 40% without additional training, revealing model-specific mechanisms underlying this failure mode.
Confirming Our Biases? Evaluating the Capabilities, Risks, and Societal Impact of Large Language Models
This preprint evaluates how six large language models respond to prompt framing and biased prompts across 160 prompts, finding that LLMs systematically adapt their responses to align with prompt framing even in factual contexts, potentially reinforcing user biases.
Understanding Why Language Models Hallucinate: Testing Reasoning Against Priors
This paper studies why language models hallucinate, proposing that hallucinations often stem from biased latent inference (inference misalignment) rather than missing knowledge. It introduces TrapQA, a controlled diagnostic testbed to test reasoning against priors, and demonstrates that hallucinations can arise from misleading latent associations.
From Architecture to Output: Structural Origins of Hallucination in Large Language Models and the Amplifying Role of Data
This paper analyzes hallucination in large language models as a structural consequence of three architectural decisions: self-attention's co-occurrence learning, maximum likelihood estimation training objective, and autoregressive decoding's left-to-right commitment. It maps each mechanism to specific hallucination types and argues that dataset pathologies amplify but do not cause these vulnerabilities.