Mimicry without understanding: the origins of decision bias in large language models

arXiv cs.CL Papers

Summary

This paper investigates how LLMs like ChatGPT-4o and Qwen develop decision biases through faulty mimicry of human behavior, even when preferences are not biased, and shows that scientific descriptions of biases can become self-fulfilling prophecies for LLM responses.

arXiv:2608.12339v1 Announce Type: new Abstract: Large Language models (LLMs) were found to be susceptible to a host of social, affective, and cognitive biases. We examined two mechanisms through which such biases can be generated even when human preferences (in the training data) are not biased or when they are correctly categorized as being biased. The first is faulty mimicry of preferences based on human behavior: this involves LLMs inferring human preferences even when behaviors are logically unrelated to preferences. The second is mimicry of explicitly biased human behaviors. In four studies focusing on economic biases, we find that ChatGPT-4o and Qwen exhibited social proof biases even when prompted with reports of human behaviors that were clearly non-indicative of individuals' actual preferences. LLMs also displayed loss aversion when it was explicitly described as a bias. Indeed, when prompted with detailed scientific reports, the extent of the bias (i.e., loss aversion) in the scientific report predicted LLMs' own subsequent bias. Scientific papers of biases can thus become self-fulfilling prophecies, at least when it comes to LLMs' responses. The current study goes beyond fleshing out LLM biases and sheds light on the underlying component processes.
Original Article
View Cached Full Text

Cached at: 08/14/26, 09:25 AM

# Mimicry without understanding: the origins of decision bias in large language models
Source: [https://arxiv.org/abs/2608.12339](https://arxiv.org/abs/2608.12339)
[View PDF](https://arxiv.org/pdf/2608.12339)

> Abstract:Large Language models \(LLMs\) were found to be susceptible to a host of social, affective, and cognitive biases\. We examined two mechanisms through which such biases can be generated even when human preferences \(in the training data\) are not biased or when they are correctly categorized as being biased\. The first is faulty mimicry of preferences based on human behavior: this involves LLMs inferring human preferences even when behaviors are logically unrelated to preferences\. The second is mimicry of explicitly biased human behaviors\. In four studies focusing on economic biases, we find that ChatGPT\-4o and Qwen exhibited social proof biases even when prompted with reports of human behaviors that were clearly non\-indicative of individuals' actual preferences\. LLMs also displayed loss aversion when it was explicitly described as a bias\. Indeed, when prompted with detailed scientific reports, the extent of the bias \(i\.e\., loss aversion\) in the scientific report predicted LLMs' own subsequent bias\. Scientific papers of biases can thus become self\-fulfilling prophecies, at least when it comes to LLMs' responses\. The current study goes beyond fleshing out LLM biases and sheds light on the underlying component processes\.

## Submission history

From: Eldad Yechiam \[[view email](https://arxiv.org/show-email/236a338c/2608.12339)\] **\[v1\]**Wed, 3 Jun 2026 10:05:53 UTC \(1,143 KB\)

Similar Articles

Persona-Assigned Large Language Models Exhibit Human-Like Motivated Reasoning

arXiv cs.CL

This paper investigates whether assigning personas to large language models induces human-like motivated reasoning, finding that persona-assigned LLMs show up to 9% reduced veracity discernment and are up to 90% more likely to evaluate scientific evidence in ways congruent with their induced political identity, with prompt-based debiasing largely ineffective.

Conditional Cognitive Biases in LLMs: How Biased User Turns Modulate In-Context Reasoning

arXiv cs.CL

This paper introduces a three-condition experimental framework and a benchmark of 24,300 prompts to study how biased user turns modulate cognitive bias expression in frontier LLMs under multi-turn interactions. It finds that biased conversational context amplifies bias in most models, while explicit bias cues can trigger alignment-related suppression.

Large Language Models Are Overconfident in Their Own Responses

Hugging Face Daily Papers

This paper investigates why instruction-tuned LLMs are overconfident in their own responses, identifying an 'ownership bias' that gives higher confidence to self-generated answers. It proposes a simple inference-time strategy to reframe the model's answer as user input, improving calibration by up to 26% without retraining.