TY - RPRT TI - Prompt perturbation and fraction facilitation sometimes strengthen Large Language Model scores AU - Mike Thelwall PY - 2025 UR - https://arxiv.org/abs/2512.01330 ID - 2512.01330 ER -