arXiv cs.AI / cs.LG / cs.CL·2d agoThe Missing Minimal Pair: Stereotype Evaluation in LLMs#llm-bias#stereotype-evaluation#minimal-pairsAI safety & security1
Hugging Face daily papers·10d agoALICE: In-context, Zero-shot, Mutual Information Estimation#mutual-information#foundation-model#zero-shot