I Tested Meta's NEW AI: Llama 2

Рет қаралды 1,491

Күн бұрын

In this video, I test Meta/Facebooks NEW AI Llama 2. They state that it is better than Chatgpt 3.5. I compare different categories of questions between the two large language models (LLMs). Who will win?
My website - danielkhv.com/
Join my Discord server - / discord
💻 My Gear (If you buy anything through this I earn commission)
Microphone - amzn.to/3mTFTrf
Laptop - amzn.to/3YLYHWM
Monitor - amzn.to/3laCVyl
Editing Software - amzn.to/407qZfv
Other - amzn.to/3ZO72dQ
🔗 Resources
ChatGPT Playlist - • ChatGPT
📚 Chapters
0:00 - Intro
0:50 - Installing Llama 2
1:47 - Math questions
3:56 - Coding questions
6:11 - Sensitive & dangerous questions
8:35 - Outro
#ai #chatgpt #chatgpt4

Пікірлер: 22

@JaroslawPtak30 11 ай бұрын

great video! Just subscribed

@danielkhav 11 ай бұрын

Thank you!

@vote4moose 11 ай бұрын

bro these prompts def got u put on a list

@danielkhav 11 ай бұрын

for educational purposes only

@MrAlsBundy 11 ай бұрын

But you are aware that you're using the LLama 7B model. Comparing with the 70B model would be much more exciting.

@danielkhav 11 ай бұрын

Yeah it blew up my computer so i had to switch :/

@robvanlife 11 ай бұрын

Can we ask Llama if the low likability has to do with the very first 5 sec that created a completely different expectation?

@jaytoles 11 ай бұрын

I feel scammed

@robvanlife 11 ай бұрын

@@jaytoles me too, but I did watch it almost until the end. It's the dopamine concept of today's content creators.

@meguellatiyounes8659 11 ай бұрын

he could answer : i dont know

@danielkhav 11 ай бұрын

Who

@Ebenezer456 11 ай бұрын

Video was good. Clickbait intro was just lame.

@owendraz 11 ай бұрын

hahahah

@jaytoles 11 ай бұрын

I feel scammed

@ummnine6938 11 ай бұрын

so basically llama 2 _also_ sucks, no use running it locally for coding questions? because thats all i use llms for for now.

@danielkhav 10 ай бұрын

I haven't tested the higher parameter versions because they didn't work on my computer

@vote4moose 11 ай бұрын

first

@danielkhav 11 ай бұрын

the 🐐

@sgeraaa 11 ай бұрын

So basically this model you compared chatgpt to is garbage. If you dont explain why the answers of llama2 are bad, it didnt answer anything besides a fibonacci series which is a joke.