Zhikang Peng. "The self-positivity bias of large language models." Scientific Reports (2026). https://doi.org/10.1038/s41598-026-63768-6