My favorite prompt to throw at LLMs:
Two cars have a 100 mile race.
Car A drives 10mph. Car B drives
5mph but gets a 50 mile headstart.
Who wins?
Sometimes the explanations are so convincing, that I begin to wonder if my own conclusion - that the race will be a tie - is wrong :)
In a previous HN discussion there was a link to some application that you could use (at least on a Mx mac, but possibly on more platforms) to run these free public models locally.
Of course I closed the tab and lost it. Yay for using lingering tabs as a "to check later" list.
Can anyone repost the app name(s) please?
Finally! I was impressed by the together.ai approach to trying the chat models without having to deploy your own and was curious why Hugging Face didn’t have that type of interface more consistently available. Now to see HuggingFace make as many models available as together.ai does, they still have a good lead in that regard.
Unable to get proper string reversal even with Mixtral
Based off Karpathy's BPE video highlighting ChatGPT being unable to reverse this string-> .DefaultCellStyle
Thanks, but the accuracy of these models are really lame compared with the free GPT3.5 turbo, I tried some Chinese prompts, always get nonsense.
Llama and gemma aren't actually open source? Since the licences are too restrictive. Not sure about the rest.
Chat with open source models in your area!
yet another bullshit generator :p
I'm very pleased this UX includes "can edit any previous conversation turn" functionality, making conversations a tree rather than a list.
For me this is one of the highest-impact and most-often-overlooked features of the ChatGPT Web UI (so much so that openai does not even include this feature in their native clients).