Researchers introduce Attention Calibration for Position-Fair Dense Retrieval, a method to counteract positional skew in dense retrieval. This method improves retrieval effectiveness and reduces position sensitivity by 43% in 16 evaluated combinations. The authors release their code on GitHub.