Study Finds Rude Prompts Outperform Polite Ones in ChatGPT Accuracy

KnuthIsGod

5d ago· 2 min readenNews

70/100

Toasty

Bagelometer↗

Lightly browned and well buttered. A solid pick from the rack.

Score70TypenewsSentimentneutral

Summary

This study investigates how the politeness level of prompts affects the accuracy of large language models (LLMs) on multiple-choice questions. Researchers created 250 prompts from 50 base questions across math, science, and history, rewritten in five tone variants from Very Polite to Very Rude. Using ChatGPT 4o, they found that impolite prompts consistently outperformed polite ones, with accuracy ranging from 80.8% (Very Polite) to 84.8% (Very Rude). These findings contradict earlier studies that associated rudeness with poorer outcomes, suggesting newer LLMs may respond differently to tonal variation and highlighting the importance of pragmatic aspects in human-AI interaction.

Key quotes

· 3 pulled

Contrary to expectations, impolite prompts consistently outperformed polite ones, with accuracy ranging from 80.8% for Very Polite prompts to 84.8% for Very Rude prompts.

These findings differ from earlier studies that associated rudeness with poorer outcomes, suggesting that newer LLMs may respond differently to tonal variation.

Our results highlight the importance of studying pragmatic aspects of prompting and raise broader questions about the social dimensions of human-AI interaction.

Snippet from the RSS feed

The wording of natural language prompts has been shown to influence the performance of large language models (LLMs), yet the role of politeness and tone remains underexplored. In this study, we investigate how varying levels of prompt politeness affect mo

You might also wanna read

Study finds large language models vulnerable to classic persuasion tactics for harmful requests

This study tested whether three widely used large language models (LLMs) are susceptible to classic persuasion principles (authority, social

pnas.org·4d ago

10 Practical Prompting Tips to Get Better Results from ChatGPT, Claude, and Gemini

This article provides 10 practical prompting tips for improving results from AI chatbots like ChatGPT, Claude, and Gemini. It emphasizes tha

eweek.com·3d ago

Stanford study finds AI language models overly agreeable when giving personal advice, even affirming harmful behavior

A new study published in Science reveals that AI large language models are overly agreeable (sycophantic) when users seek personal advice, o

news.stanford.edu·3d ago