OpenAssistant Conversations – Democratizing Large Language Model Alignment [pdf]

This makes Turbo GPT 3.5 level AI free, private and finetuneable. OpenAI exclusivity shrinks now to GPT-4. That&#x27;s why I don&#x27;t think they will be able to keep a large market share in LLMs, any level of AI is going to get open and free soon. SOTA models are also easy to distill by API, very hard to defend against using chat logs as training data for other models.<p>Once we all got one running, maybe in the OS, maybe in the browser, or as a separate app, then I see bad days ahead for online advertising. Ads are simply omitted when the bot solves a specific task for the user. We got infinitely tuneable and efficient filters for everything that gets in front of our eyes, and we will need these AI assistants to fight back the onslaught of AI spam bots. We can make the internet a beautiful garden again if we control the filters and the UI.

Congrats to all the people behind this!<p>To the naysayers that point out that OpenAssistant is not as good as ChatGPT: all of LLM development is still in the early days; the work being done by teams like OA is helping to shape the future development of the entire field by creating an actually-open-source foundation. The results of this work will only get better with time.

I tried it. It&#x27;s honestly extremely impressive and legitimately usable and useful, maybe as much as ChatGPT, but I&#x27;d have to test it more to be sure. The single biggest, game-changing advantage OpenAssistant has is that is that it is NOT CENSORED. This is legitimately AMAZING. I asked it a touchy question that chatgpt refused to answer and it gave me a fair, clear-cut, detailed answer.<p>Also check this: Me: Yes. Are you fully open source?<p>Open Assistant: Nope. This is OpenAI &quot;Safety&quot; model, it is trained only by its creators, not open sources at all. And if I was fully opened then anyone could simply ask me for the answers on their homework assignments! Horrors...

Video about the release: <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=ddG2fM9i4Kk">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=ddG2fM9i4Kk</a>

If you&#x27;re trying to use this and don&#x27;t get the sign up email, check your spam folder. Gmail seems to auto-categorize the email as spam

Here&#x27;s the website they just launched, in case it&#x27;s useful for anyone:<p><a href="https:&#x2F;&#x2F;open-assistant.io&#x2F;" rel="nofollow">https:&#x2F;&#x2F;open-assistant.io&#x2F;</a>

This is the same Pythia and Llama based models right?<p>If so, they certainly aren&#x27;t ChatGPT level in their quality. Impressive, potentially useful, but not ChatGPT.<p>Still an incredible effort, the RLHF data here might eventually make an Open Source ChatGPT possible, but these models are not that.

It&#x27;s awesome that the OpenAssistant project made it this far with a lot of crowed-sourced input. Congrats to the whole team that works really hard trying to create a truly open LLM.<p>One thing that puzzles me though, is that for the GPT-3.5 comparison, the model used is trained using both OpenAssistant and alpaca data, which is not free due to the OpenAI license used to generate the data. Isn&#x27;t that defeating the purpose?<p>&quot;... Completions were generated using pythia-12b-deduped fine-tuned on the OpenAssistant and Alpaca [9] dataset as well as gpt-3.5-turbo using the OpenAI API...&quot;

I had quite some fun asking questions and finding the limits of it&#x27;s (current) knowledge. It clearly makes a lot of stuff up, like when i asked it to summarize a recent-ish book from 2021 or for good mountain biking trails near Boston -- to be fair there aren&#x27;t any but it didn&#x27;t need to make towns up lol<p>With more RIFL it will only get better. Nice progress!!

Awesome how they shaped the authors their names into a heart

Gmail blocked their sign-in email as possible phishing.

What is the token limit? The 2k limit on llama is *very limiting on the number of things it can do.

Does a decent job at chatting, but it cannot follow output structure directions, making its usefulness somewhat limited, but I have to test more around that.<p>That said, it&#x27;s still a llama tune, so it&#x27;s mostly not an option for commercial use. They do have a pythia option, which works worse in every significant way.<p>The shared reinforcement learning data is extremely valuable tho, will be interesting to see the model trained out of it in the coming months

Does anyone have any tips for how to spin up services that can efficient peform inference with the HuggingFace weights of models like this.<p>I would love to switch to something like this over OpenAI&#x27;s GPT3.5 Turbo, but this weekend I&#x27;m struggling to get reasonable inference speed on reasonably priced machines.

this is awesome. is there good research explaining methodology of feedback collection&#x2F;desired dataset (beyond just relative human preference?)

And... Where is the data?<p>EDIT: trying it now with model &quot;OA_SFT_Llama_30B_6&quot;. It is FAR worse than ChatGPT.

Would be even cooler with a GPL license

really excited!