OpenAI 推出 LifeVal 基准，测试 ChatGPT 处理生活杂务的能力
OpenAI 搞了个新基准 LifeVal，专门测 ChatGPT 能不能帮你办杂事，GPT-6 Astra 十六项里十项满分，成绩单都列出来了。
OpenAI 基于 3,995 人时间使用调研和生活琐事构建了 LifeVal 测试，覆盖 16 项真实生活任务。GPT-6 Astra 在其中 10 项任务上拿到 100% 完成率，整体平均得分 80.6%。单项成绩包括：找回订婚戒指匹配 100%、退订邮件列表 100%、德州食品补助文书 98.7%、为糖尿病家长制定饮食计划 90.4%、核对 LSAT 答案 88.7%。
OpenAI says it wants ChatGPT handling your life admin, so it built LifeVal, a test of real chores picked using studies of how people spend their time and a 3,995-person poll. GPT-6 Astra hit 100% on 10 of the 16 tasks in its results table. A few, with its "best recorded" score on each: - Finding a match for a lost engagement ring: 100% - Quitting email lists: 100% - Texas food-aid paperwork: 98.7% - Diet plan for a diabetic parent: 90.4% - Checking LSAT answers: 88.7% Its average across the whole test is 80.6%. Your browser does not support the video tag. 🔗 View on Twitter 💬 9 🔄 0 ❤️ 12 👀 2508 📊 9 ⚡
