1
Localizing Anchoring Pathways in Language Models
探寻语言模型内部语义锚定机制，精准定位关键路径，推动模型可解释性研究。
arXiv:2606.12818v1 Announce Type: cross Abstract: Irrelevant numbers in a prompt can shift language model judgments, producing anchoring effects in nu…
探寻语言模型内部语义锚定机制，精准定位关键路径，推动模型可解释性研究。
arXiv:2606.12818v1 Announce Type: cross Abstract: Irrelevant numbers in a prompt can shift language model judgments, producing anchoring effects in nu…