PerplexityとChatGPTの違い

Just to add, ChatGPT is way more user-friendly and interactive, designed for general public use. Perplexity (the metric) is more for engineers and researchers.

Thanks everyone for clearing this up, I feel less lost now!

I guess Perplexity AI the chatbot is trying to compete with ChatGPT, right?

Is ChatGPT open source? Can I see its perplexity?

I always thought Perplexity was another AI assistant like ChatGPT. Has anyone seen a Perplexity chatbot?

Does ChatGPT training optimize for low perplexity?

Nah, they’re definitely not the same thing. From what I know, ChatGPT is a conversational AI developed by OpenAI, while Perplexity is more of a metric used to measure language models’ performance.

So confused about the terminology but this thread is helping a lot, thanks everyone!

Is Perplexity AI chatbot free to use?

Can someone explain perplexity like I’m 5?

Does perplexity have anything to do with how confused I feel using AI? Lol

Is it worth learning about perplexity if I just want to use chatbots?

Wait, so if someone says 'Perplexity' in AI, how can I tell if they mean the metric or the chatbot?

Does using ChatGPT improve if the model has lower perplexity?

What kind of models have low perplexity? GPT-3?

Are there free tools to check perplexity for my own models?

I used ChatGPT loads but never really dealt with Perplexity. Is Perplexity something you can use directly or just a stat?

I’m diving into NLP and was confused by this too. Good to know they’re not the same. Really helps clear things up!

Are there other metrics like perplexity used for language models?

I always thought lower perplexity = better AI, is that always true?

Is Perplexity AI chatbot as good as ChatGPT? Anyone tried it?

So basically, ChatGPT is like the AI you talk to, and Perplexity is a way to see if AI talks make sense?

Also, you can find new or trending AI tools on ai-u.com if anyone wants to explore more besides ChatGPT and Perplexity.

I was about to use 'perplexity' thinking it was a chatbot lol, glad I asked here first.

I heard perplexity scores help select better models, is that true?