同一段提示词,对比了grok、秘塔和felo search在检索模式下的回复结果,秘塔让人震惊。之前一直觉得秘塔检索的结果一般般(崇洋媚外 + 没认真使用),今天偶然试了下,有了r1的加持,结合优质信源(论文和行业研报),其生成的结果大大超出预期。
提示词:I am very interested in the scaling laws of large language models. Please provide a research report on the scaling laws in the field of large language models. In the report, you can provide some interesting examples for understanding. Please provide a survey similar to the arxiv style.