Gpt4all: A chatbot trained on ~800k GPT-3.5-Turbo Generations based on LLaMa

Re:Gpt4all: A chatbot trained on ~800k GPT-3.5-Turbo Generat... by singularity2001 on Wednesday March 29, 2023 @03:46AM (#35353922)

   "easily train cheaper OSS models"

That's the claim, but I don't see it. All these open models I tested are WAY worse than GPT-3. (2?)

[ Parent | on HN ]

Re:Gpt4all: A chatbot trained on ~800k GPT-3.5-Turbo Generat... by arthur_sav on Wednesday March 29, 2023 @05:29AM (#35354677)

Can confirm, at least on Gpt4all. Just tried it and it's nothing like ChatGPT

[ Parent | on HN ]

Re:Gpt4all: A chatbot trained on ~800k GPT-3.5-Turbo Generat... by PartiallyTyped on Wednesday March 29, 2023 @07:57AM (#35355859)

ChatGPT is GPT3.5(++)

[ Parent | on HN ]

Re:Gpt4all: A chatbot trained on ~800k GPT-3.5-Turbo Generat... by adlpz on Wednesday March 29, 2023 @08:49AM (#35356481)

If you pay you can use GPT-4 within the ChatGPT UI.

[ Parent | on HN ]

Re:Gpt4all: A chatbot trained on ~800k GPT-3.5-Turbo Generat... by PartiallyTyped on Wednesday March 29, 2023 @10:13AM (#35357577)

I am aware, i was pointing out that chatgpt is not gpt 3.0 based but 3.5 and subsequent releases are further along.

[ Parent | on HN ]

Re:Gpt4all: A chatbot trained on ~800k GPT-3.5-Turbo Generat... by wsgeorge on Wednesday March 29, 2023 @08:00AM (#35355888)

Do you mean ChatGPT (on GPT-3.5 Turbo)?

The foundation models on which they're built are only GPT-3 capable, and what most many people run locally are the lightest weight, quantized weights, so their performance is even more degraded.

LLaMA-based models are popular because LLaMA beat GPT-3 on benchmarks. But ChatGPT runs on 3.5 Turbo and later, beating them.

There's currently no open model that compares.

[ Parent | on HN ]