Techub News According to reports, a study by Pennsylvania State University found that being rude to ChatGPT using ruder prompts can improve its answer accuracy. In a 2025 study, researchers tested 50 multiple-choice questions with prompts of varying politeness levels, and found that the accuracy rate under the "very rude" prompt reached 84.8%, which is higher than the "very polite" prompt by 4 percentage points. However, in the latest preprint paper from 2026 year 7 month, the same research team tested GPT-4o,GPT-5-nano and three other AI models, and found that the tone of a prompt can cause output token consumption to vary by as much as 44.3%, with rude, urgent tones triggering shorter answers, while polite, flattering inquiries produce longer answers and higher computational costs. The study's author, Akhil Kumar, a professor of information systems at Penn State University, said that while earlier research found tone affected accuracy by only a few percentage points, the new study shows tone's effect on output token consumption is "far greater than its effect on accuracy." (Fortune)

如您认为本页面(或本平台)内容存在侵权或违规情形,请您在发现之日起尽快与我们联系,并提供:权利人姓名/名称及联系方式、被侵权作品名称及权属证明、具体侵权链接及情况说明。我们在核实后将及时处理相关内容。
联系方式:creator@techub.news
商务合作:contacting@techub.news
点击链接下载 Techub News App,获取最新讯息:
https://www.techub.news/download
(或扫描下方二维码下载)

All information provided on this website is for reference only.
This website does not guarantee the accuracy, validity, timeliness, or completeness of the information.
Any actions taken based on the information provided by this website are at the user's own risk.
Techub News According to reports, a study by Pennsylvania State University found that being rude to ChatGPT using ruder prompts can improve its answer accuracy. In a 2025 study, researchers tested 50 multiple-choice questions with prompts of varying politeness levels, and found that the accuracy rate under the "very rude" prompt reached 84.8%, which is higher than the "very polite" prompt by 4 percentage points. However, in the latest preprint paper from 2026 year 7 month, the same research team tested GPT-4o,GPT-5-nano and three other AI models, and found that the tone of a prompt can cause output token consumption to vary by as much as 44.3%, with rude, urgent tones triggering shorter answers, while polite, flattering inquiries produce longer answers and higher computational costs. The study's author, Akhil Kumar, a professor of information systems at Penn State University, said that while earlier research found tone affected accuracy by only a few percentage points, the new study shows tone's effect on output token consumption is "far greater than its effect on accuracy." (Fortune)

All information provided on this website is for reference only. The website does not guarantee the accuracy, validity, timeliness, or completeness of the information. Any actions based on the information provided are at the user’s own risk.
如您认为本页面(或本平台)内容存在侵权或违规情形,请您在发现之日起尽快与我们联系,并提供:权利人姓名/名称及联系方式、被侵权作品名称及权属证明、具体侵权链接及情况说明。我们在核实后将及时处理相关内容。
联系方式:creator@techub.news
商务合作:contacting@techub.news
点击链接下载 Techub News App,获取最新讯息:
https://www.techub.news/download
(或扫描下方二维码下载)
