LLM GPT-4.1 tested for human bias in random number generation

Researchers investigated whether large language models like GPT-4.1 exhibit human biases in generating random numbers. When asked to pick a number between 1 and 100, humans tend to favor certain numbers and avoid others, deviating from a uniform distribution. This study tested GPT-4.1 with the same prompt 10,000 times to compare its output distribution against a true random generator. AI

IMPACT Examines whether LLMs inherit human cognitive biases, impacting their use in applications requiring true randomness.

RANK_REASON The cluster describes a research study analyzing the output of an LLM against human behavior. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Mastodon — fosstodon.org →

• GPT-4.1
• humans

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →