Meaningless is better: Hashing bias-inducing words in LLM prompts improves performance in logical reasoning and statistical learning

• Our new masking technique replaces biased words with referenceable identifiers, significantly improving Linda problem & itemset extraction. • LLMs tested on an adapted "Linda problem" show human-like susceptibility to the representativeness heuristic and "and" fallacy. • Experimental results show tabular formatting, rather than free text, reduces bias in LLMs, which may also apply to humans. • LLMs successfully extracted frequent itemsets from small data without code, showing broad implications for algorithmic rule learning. This paper introduces a novel method, referred to as "hashing", which involves masking potentially bias-inducing words in large language models (LLMs) with hash-like meaningless identifiers to reduce cognitive biases and reliance on external knowledge. The method was tested across four sets of experiments involving a total of 640 prompts. Statistical analysis using chi-square tests showed significant improvements in all tested scenarios, which covered the Llama, ChatGPT, Copilot, Gemini and Mixtral models. In the first experiment, hashing decreased the conjunction fallacy rate in a modified version of the “Linda” problem aimed at evaluating susceptibility to cognitive biases. In the second experiment, it improved LLM results on the frequent itemset extraction task. In the third experiment, we found hashing is also effective when the Linda problem is presented in a tabular format rather than text, indicating that the technique works across various input representations. In the fourth experiment, consistent with psychological literature showing that step-by-step reasoning suppresses biases, we compared hashing's effectiveness with Chain of Thought (CoT) LLM models, finding that CoT also suppressed biases in LLMs. Overall, the proposed hashing method was shown to improve bias reduction and reduce the unwanted incorporation of external knowledge. Despite bias reduction, reduction in hallucination rates was inconsistent across LLM model types. These findings suggest that masking bias-inducing terms can improve LLM performance, although its effectiveness is model- and task-dependent.