[{"data":1,"prerenderedAt":81},["ShallowReactive",2],{"tool-c-eval":3},{"tool":4,"relatedTools":43},{"id":5,"name":6,"slug":7,"tagline":8,"description":9,"websiteUrl":10,"logoText":11,"pricing":12,"editorScore":13,"clickCount":14,"featured":15,"latest":15,"updatedAt":16,"category":17,"tags":26,"overview":28,"highlights":29,"useCases":38},"fd177d677a75e068b5b3a441fd1a4626","C-Eval","c-eval","一个全面的中文基础模型评估套件","C-Eval是一个适用于大语言模型的多层次多学科中文评估套件，由上海交通大学、清华大学和爱丁堡大学研究人员在2023年5月份联合推出，包含了13948个多项选择题，涵盖了52个不同的学科和四个难度级别，用以评测大模型中文理解能力。","https:\u002F\u002Fcevalbenchmark.com\u002Findex_zh.html#home_zh?utm_source=ai-bot.cn","CEV","官网可查看",98,12,false,"2026-04-15T04:26:34+08:00",{"id":18,"name":19,"slug":20,"description":21,"iconLabel":22,"accentColor":23,"toolCount":24,"sortOrder":25},"d6d303a9f41677e5f1a5f9b12c520225","智能体与自动化","agents","适合搭建业务工作流、插件编排和多步骤自动化，用来把 AI 接进真实流程。","AUTO","#6C7A2D",216,7,[27],"AI模型评测","C-Eval是适用于大语言模型的多层次多学科中文评估套件，由上海交通大学、清华大学和爱丁堡大学研究人员在2023年5月份联合推出，包含13948个多项选择题，涵盖52个不同的学科和四个难度级别，用在评测大模型中文理解能力。通过零样本（zero-shot）和少样本（few-shot）测试，C-Eval 能评估模型在未见过的任务上的适应性和泛化能力。 多学科覆盖 ：C-Eval 包含 52 个不同学科的题目，涵盖 STEM、社会科学、人文科学等多个领域，全面评估语言模型的知识储备。",[30,31,32,33,34,35,36,37],"多学科覆盖","多层次难度分级","量化评估与标准化测试","数据下载","Hugging Face 下载：","或者直接下载 ZIP 文件并解压：","选择评估模式","零样本（Zero-shot）",[39,40,41,42],"13948个多项选择题","52 个不同学科的题目","社会科学","人文科学等多个领域",[44,57,68],{"id":45,"name":46,"slug":47,"tagline":48,"description":49,"websiteUrl":50,"logoText":51,"pricing":12,"editorScore":13,"clickCount":52,"featured":53,"latest":53,"updatedAt":16,"category":54,"tags":55},"a638bfd4cc081d0294d7d12975a727dd","GPT-4","gpt-4","OpenAI旗下最新的GPT-4模型","GPT-4（Generative Pre-trained Transformer）是 OpenAI 开发的自然语言处理模型 GPT 家族中的第四个版本，该模型依靠强大的神经网络来理解和生成类似人类的语言。","https:\u002F\u002Fcn.bing.com\u002Fsearch?q=GPT-4&ensearch=1","GPT",2356,true,{"id":18,"name":19,"slug":20,"description":21,"iconLabel":22,"accentColor":23,"toolCount":24,"sortOrder":25},[56],"AI训练模型",{"id":58,"name":59,"slug":60,"tagline":61,"description":62,"websiteUrl":63,"logoText":64,"pricing":12,"editorScore":13,"clickCount":65,"featured":53,"latest":53,"updatedAt":16,"category":66,"tags":67},"a23e89d9e4203bd1c683c41e837f370e","LMArena","lmarena","AI模型评估平台","LMArena是加州大学伯克利分校推出的创新AI模型评估平台，基于让用户对不同AI模型的回答进行匿名投票，衡量模型的表现。用户输入问题后，平台提供两个模型的回答，用户根据偏好选择更优答案，投票结果直接塑造公共排行榜。","https:\u002F\u002Flmarena.ai\u002F?utm_source=ai-bot.cn","LMA",681,{"id":18,"name":19,"slug":20,"description":21,"iconLabel":22,"accentColor":23,"toolCount":24,"sortOrder":25},[27],{"id":69,"name":70,"slug":71,"tagline":72,"description":73,"websiteUrl":74,"logoText":75,"pricing":76,"editorScore":13,"clickCount":77,"featured":53,"latest":53,"updatedAt":16,"category":78,"tags":79},"32bb7d109995125694dc4a175a840117","Coze","coze","海量AI智能体免费用，一键复制同款","Coze是字节跳动推出的AI办公助手，能通过AI技术简化工作流程，提升工作效率。通过Agent Skills技能系统和Agent Plan长期计划功能，实现从单一工具到智能工作伙伴的转变。","https:\u002F\u002Fwww.coze.cn\u002Foverview?utm_medium=daohang&utm_source=aibot&utm_term=49156769m","COZ","免费",316,{"id":18,"name":19,"slug":20,"description":21,"iconLabel":22,"accentColor":23,"toolCount":24,"sortOrder":25},[80],"AI开发平台",1776262187806]