Naive LLM judges are inconsistent. Run the same poem through twice and you get different scores (obviously, due to sampling). But lowering the temperature also doesn’t help much, as that’s only one of many technical issues. So, I developed a full scoring system, based on details on the logits outputs. It can get remarkably tricky. Think about a score from 1-10:
newrepublic.com
。使用 WeChat 網頁版对此有专业解读
从这个渡口登舟远行，唐诗如同一条星河。陈寅恪认为中国诗歌区别于外国诗歌最根本者，在“与历史之关系”：“中国诗虽短，却包括时间、人事、地理三点”。时间、人事、地理，使得中国的文学总是锚定大地和人间，这是最为悠远和辽阔的现实主义。沿着这条星河往前驶行，你会发现，唐诗的永恒魅力，不只在于其辞藻与意境的华美，更在于它承载着一代代中国人健卓顽韧的精神力量与生命咏叹。
�@e�X�|�[�c�ƊE�Ŕ|���Ă����m���������\�����Ă��p�r�ɉ�����BTO���f���̓W�J���s���Ă����\���B�ڋq�̗v�]�ɉ������t���I�[�_�[���C�h���p�ӂ����Ƃ��Ă����B，更多细节参见谷歌
Open up the app and connect to a server in a location with access
Collaboration between Microsoft, Google, Mozilla, Bloomberg, Igalia, Boa, and many independent contributors，这一点在超级权重中也有详细论述