Home /permanent

Position Bias

Position Bias is a phenomenon in LLM-based recommender systems in which the model's recommendations are disproportionately influenced by the order of candidate items in the context, rather than solely by their relevance to the user's preferences (Bito et al., 2025).

References

Ethan Bito, Yongli Ren, and Estrid He. Evaluating Position Bias in Large Language Model Recommendations. August 2025. arXiv:2508.02020, doi:10.48550/arXiv.2508.02020.