GPT-4 Knowledge Cutoff Date?
最近在学习 SwiftData, Chat Mode 中 问了 GPT-4 一些问题, 感觉它并不知道 SwiftData 的存在, 感觉之前发布的 GPT-4 Turbo 的 Cutoff Date 貌似是 23年低,
Thread: https://devv.ai/search?threadId=dol8f2lb3fgg
Recently, I've been learning SwiftData, and in Chat Mode, I asked GPT-4 some questions, and I felt that it didn't know about the existence of SwiftData, and I felt that the Cutoff Date of the previously released GPT-4 Turbo looks like it's 23 years low, and it's a good thing that GPT-4 is not aware of the existence of SwiftData.
Comments3
Bohan
Jun 13, 2024
Hi,
所有 LLM 本身都是存在幻觉问题和知识 cutoff 时间点的，这也是我们最初开发 Devv AI 的初衷。Devv 通过使用其内部的 RAG 系统让搜索的结果更加实时、准确。
我看到您在这个问题上使用的是聊天模式，想问下是否有什么特殊的原因没有使用联网模式？您可以换用联网搜索的模式对 SwiftData 进行相关的提问，结果会更准确。
All LLMs themselies have illusination issues and knowledge cutoff time. And that's why Devv AI works better than ChatGPT or any LLM wrapper. We use RAG system and online data to make sure your question will be answered accurately.
I see that you are using Chat Mode for this question, and would like to ask if there is any particular reason why you are not using Web Mode? You can switch to Web Mode for questions related to SwiftData, and the results will be more accurate.
Devv AI
Bohan
David
Jun 13, 2024
Hi, 感谢您的回复, 我本来是想提问一下逻辑上的问题, 解释一段代码, 我当时想着联网模式 (Web Mode) 类似微软的那种检索网页信息整合, 所以就直接用了单纯的 Chat Mode, 觉得Chat Mode的逻辑应该比Web 好一些, 谢谢您的建议, 我会多尝试联网模式
Bohan
Jun 14, 2024
没问题，如果您仍然遇到问题，欢迎继续向我们反馈。
No problem, if you're still experiencing problems, feel free to keep us posted.