More is More: Addition Bias in Large Language Models

Luca Santagata, Cristiano De Nobili

In this paper, we investigate the presence of additive bias in Large Language Models (LLMs), drawing a parallel to the cognitive bias observed in humans where individuals tend to f…