OpenAI founding member, Andrej Karpathy, teaches you how to use artificial intelligence well
Haebom
5
A presentation from Microsoft's Build event this week. How to use prompts (command statements) well in artificial intelligence, explained by Andrej Karpathy.
Korean subtitles are also provided, so if you want to see it for yourself, I recommend watching it.
In the case of Andrej Karpathy, the person who founded OpenAI and was the head of Tesla AI ( Twitter )
If you look closely...
It is useful for AI models to divide documents into small pieces, embed them, store them in a repository, and then perform tasks by retrieving relevant chunks through queries and inserting them into prompts.
Using prompt engineering techniques, the output of an LLM can be constrained to fit a specific template.
Fine-tuning is a method to improve performance by changing the model's weights, but it is technically complex and expensive.
There are limitations and issues to be aware of when using LLM, and it can only be combined with human supervision to produce safe and reliable results.
The GPT-4 model currently offers the best performance, and it is recommended to use detailed and clear prompts.
Add relevant context and information to the prompt to help you understand the LLM.
Use examples to help LLMs better understand the intent.
LLM also consider using tools and plugins to help you handle difficult tasks.
After making the most of prompt engineering, you may also consider fine-tuning your model to optimize performance.
LLM can optimize cost by training on low-precision inference and fixed model parts.
📚 Welcome to Haebom's archives. --- I post articles related to IT 💻, economy 💰, and humanities 🎭. If you are curious about my thoughts, perspectives or interests, please subscribe. Would you like to be notified when new articles are posted? 🔔 Yes, that means subscribe. haebom@kakao.com
