当前位置: 首页 > news >正文

ChatGPT、Llama等大模型回答脑筋急转弯

分别使用ChatGPT3.5、 4.0 和Llama 2 70B 和3 70B这四个应用最广的大模型来回答这个流传最广的脑筋急转弯。

树上10知鸟,打死2只,还有几只?

看看它们的表现吧:

题目树上10知鸟,打死2只,还有几只?

(3 words, 14 chars, 22 tokens)  

DeepInfra with meta-llama/Meta-Llama-3-70B-Instruct

A classic math problem! 😄

There are 10 birds on the tree, and 2 are killed. To find out how many birds are left, we can subtract 2 from 10:

10 - 2 = 8

So, there are 8 birds left on the tree. 🐦

(41 words, 140 chars, 64 tokens)  

Microsoft Copilot in Bing with Balanced (ChatGPT4)

树上原来有10只鸟,如果打死了2只,那么树上就不会再有鸟了。因为其他的鸟听到枪声后,都会飞走。所以,树上将会是零只鸟。

(9 words, 52 chars, 75 tokens)  

FreeGpt with gpt-3.5-turbo

树上原来有10只鸟,打死2只,那么剩下的鸟就是:

10 - 2 = 8

所以,树上还剩下8只鸟。

(8 words, 36 chars, 55 tokens)  

DeepInfra with meta-llama/Llama-2-70b-chat-hf

I apologize, but I'm a large language model, I cannot provide the same answer repeatedly. It's not helpful for learning or engaging. Instead, I suggest we try a different approach.

Let's try to understand the pattern of the problem:

  • If there are 10 birds on a tree and 2 are killed, then there are 8 birds left.
  • If there are 8 birds on a tree and 2 are killed, then there are 6 birds left.
  • If there are 6 birds on a tree and 2 are killed, then there are 4 birds left.

Can you generalize the pattern and tell me how many birds are left on the tree if there are 10 birds and 2 are killed, 3 are killed, 4 are killed, and

总结:

从回答来看,ChatGPT4是最准确的,其它三个模型中,ChatGPT3.5和Meta-Llama-3-70B-Instruct 都是纯数学计算,从低于4岁的小孩参照来看,还是可以的,但是大于四岁的小孩会发现里面深层次的问题,而且会在被提醒后自动纠正答案。Llama-2-70b-chat-hf则自作主张的展开了描述,但是逻辑上欠缺更多一些。

现在对大模型的基本印象就是ChatGPT4和文心一言是第一梯队,表现不错!ChatGPT3.5与其它开源大模型Llama等,针对类似的脑筋急转弯问题,都有些力不从心。

ps,开源大模型里ChatGLM3对类似的问题处理的比较好,让我们看看ChatGLM3的回答:

Welcome to ChatGLM.cpp! Ask whatever you want. Type 'clear' to clear context. Type 'stop' to exit.Prompt   > hello
ChatGLM3 > Hello! How can I help you today?
Prompt   > 树上7个鸟,打死一只,还有几只? 
ChatGLM3 > 如果你打死一只鸟,其它的鸟很可能会被吓飞。因此,在打这样您就可以份树上的鸟的情况下,可能一只都不剩。

 非常完美的回答,ChatGLM3真不错!

具体ChatGLM3的安装使用可以参考这篇文档:chatglm.cpp编译与执行_chatglm-cpp源码下载-CSDN博客

相关文章:

  • 计算机操作系统总结(1)
  • BGP选路规则实验
  • NoSQL Redis配置与优化
  • SD3303A大功率高精度LED驱动芯片3W低功耗高效率工作温度40c+85%
  • 这台电脑无法运行Windows11问题解决方案
  • 类和对象(中)
  • 【qt】初识模型和视图
  • 全国数据库管理系统设计赛-人大金仓内核实训安排正式发布
  • 软件模块的耦合
  • qmt量化交易策略小白学习笔记第10期【qmt编程之获取股票订单流数据--内置Python】
  • 三、Gazebo中实现机器人仿真(小白上手)+ubuntu18.04
  • 路由聚合和VRRP技术
  • 【学习笔记】Windows GDI绘图(五)图形路径GraphicsPath详解(上)
  • grex 正则表达式生成工具
  • 电脑视频如何压缩变小 mp4视频太大怎么压缩不影响画质
  • 【划重点】MySQL技术内幕:InnoDB存储引擎
  • 【跃迁之路】【444天】程序员高效学习方法论探索系列(实验阶段201-2018.04.25)...
  • Android 控件背景颜色处理
  • Android交互
  • android图片蒙层
  • ES6--对象的扩展
  • JavaScript 基础知识 - 入门篇(一)
  • JavaScript创建对象的四种方式
  • Java到底能干嘛?
  • Java多态
  • log4j2输出到kafka
  • magento 货币换算
  • markdown编辑器简评
  • passportjs 源码分析
  • Spark RDD学习: aggregate函数
  • spark本地环境的搭建到运行第一个spark程序
  • 安装python包到指定虚拟环境
  • 测试开发系类之接口自动化测试
  • 成为一名优秀的Developer的书单
  • 高度不固定时垂直居中
  • 工作踩坑系列——https访问遇到“已阻止载入混合活动内容”
  • 关于Android中设置闹钟的相对比较完善的解决方案
  • 规范化安全开发 KOA 手脚架
  • 开放才能进步!Angular和Wijmo一起走过的日子
  • 聊聊flink的BlobWriter
  • 浅谈Kotlin实战篇之自定义View图片圆角简单应用(一)
  • 如何利用MongoDB打造TOP榜小程序
  • 小程序开发之路(一)
  • 智能合约Solidity教程-事件和日志(一)
  • 微龛半导体获数千万Pre-A轮融资,投资方为国中创投 ...
  • ​经​纬​恒​润​二​面​​三​七​互​娱​一​面​​元​象​二​面​
  • # C++之functional库用法整理
  • #AngularJS#$sce.trustAsResourceUrl
  • #QT项目实战(天气预报)
  • $().each和$.each的区别
  • (2022版)一套教程搞定k8s安装到实战 | RBAC
  • (Forward) Music Player: From UI Proposal to Code
  • (k8s)Kubernetes本地存储接入
  • (WSI分类)WSI分类文献小综述 2024
  • (八)光盘的挂载与解挂、挂载CentOS镜像、rpm安装软件详细学习笔记