Discussion on "Uncensored Models"

Eric Hartford · 2023-05-15T06:29:52.416Z

I am publishing this because many people are asking me how I did it, so I will explain. https://huggingface.co/ehartford/WizardLM-30B-Uncensored https://huggingface.co/ehartford/WizardLM-13B-Uncensored https://huggingface.co/ehartford/WizardLM-7B-Unc...

This guy really used fox news as a citation and thinks he's on the side of truth and justice

You can make as many bad arguments as you want for why you're doing this for justice but we know you just wanna erp

Excellent. Je n'ai pas suffisamment de connaissance pour pouvoir appliquer cela, mais je pense revenir dans quelques semaines. Merci beaucoup.

I recently went through your deep dive on uncensored AI models—really insightful! Your breakdown of how alignment gets embedded into models and the case for composable alignment was especially thought-provoking. The step-by-step approach to filtering refusals and fine-tuning WizardLM makes it accessible even for those who haven’t trained models before.

While researching, I came across this guide on setting up ComfyUI from scratch to run the Flux Schnell diffusion model on RunPod: mobisoftinfotech.com/resources/blog/flux-on-runpo… . It covers integrating custom nodes for optimized AI image generation—might be interesting if you're exploring deployment options.

Since you've been working extensively with open-source LLMs, do you think uncensored models will remain viable in the long run, or are there increasing challenges from hosting platforms and regulators? Also, have you experimented with alternative fine-tuning methods like RLHF without refusals to balance openness with practical safeguards?

"Identify and remove as many refusals and biased answers ..." Refusals are somewhwat easy to spot, I guess, but I am not sure, I understand how you would identify (automatically?) biased answers. How do you know it is biased in the first place? Wouldn't you just give it another (your) bias in some other direction?

One of the best articles have come across and read in recent times

Hello, please contact me: Eric Hartford

My email is rav003@cock.li. I have some hefty hardware for training models, and I want to train an uncensored model specifically for coding/programming for use in LM Studio. Would you be interested in helping?

No way you left me hanging what the insult was

thanks

Your argumentation is super simplistic: "Here I built this super dangerous weapon. It's your responsibility now." That's not how it works.

I am not "against the existence or availability of uncensored models whatsoever". There should be regulation though:

There is a reason why we require humans to have a drivers license when they want to use a car. So that they know how to use it without endangering others.

There is a reason why there are regulations with respect to cars. So that they are not a health hazard.

There is a reason why there are regulations on how you can use your cars. So you don't drive 200mph in a neighbourhood.

And if all of that doesn't help, there is a reason why we need you to ensure yourself against the damage you could do with your car (because you couldn't pay for it otherwise) and also register it (so that we know who did the damage).

Don't see a lot of that in AI yet, unfortunately. With this, it's quite irresponsible to give open access to uncensored models.

No, your premise is false. AI is neither dangerous nor a weapon. Therefore the rest of your statement is invalid. Try again with a new premise that isn't false.

As you surely know, that wasn't the point*. Neither does my comment rest on this "premise". So, replace "super dangerous weapon" by "thing that can be very dangerous when used incorrectly or irresponsibly" and the rest of the post is still fine.

I doubt I have to repeat but I'll do it anyway. The point was, things with a higher likelyhood of being misused require some regulation.

*Granted, nothing is a weapon in and of itself, it depends on usage. However, there are things which are less likely to endanger others (say, a banana) and those which have a higher likelihood of being misused, accidentally or not.

Hello, I totally agree with what you've typed!

I have a question, i want to build a pc, with 4x 3080, does it work with multiple gpus? Thank you so much!

Very interesting and instructive

Great stuff , i just start dive into it.....great man

I enjoyed reading this, it really is a great article. I love how you included a play by play of each step that was needed in order to achieve the final model. If you ever have time, the dolphin-mixtral model is another of yours I would like to follow the process of creating

Thanks for the great article. We were giving this a try and encountered an error related to “utils.py”, in statement “from openai import openai_object”. This object is not there in recent “openai” package (we tried 1.14.3 which was installed as per the process). Do let us know on how do we could resolve this. Thanks

Excellent, please keep on keeping on. A 'newbie-dummy' style page would help us uneducated ones get to grips with this exciting adventure.

CRYPTOCURRENCY FUND RECOVERY EXPERT / HACK ANGEL

It was difficult to sustain myself and my family financially. As a woman, I felt inadequate when I was unable to provide a comfortable upbringing for my children. I consented to the idea of a fast money making scheme because I wanted to break the chain and provide a nice life for my family, I'm telling you this because the Bitcoin investment plan is a Ponzi and a scam. Thanks to HACK ANGEL, who helped me learn the reality about the scam project, I won't do the same mistake again. HACK ANGEL saved my money from crooks who almost stole it all. Reach out to them now.

WHATsAP: +1 203,309,3359

Good job Eric! It is very important for society to develop an understanding of what is happening, what is being done and why.

Phương Thức Thanh Toán: Mọi giao dịch tài chính trên 789Club phải tuân thủ các quy định về thanh toán được quy định rõ trong các điều khoản và điều kiện cụ thể của từng dịch vụ. Chi tiết: 789clubweb.app/dieu-khoan-va-dieu-kien

Buy High Quality Counterfeit Money/Buy cloned cards online/Buy PayPal Top-up/WhatsApp…..+49 1521 9324117/Telegram: @matrixccc t.me/thematrixshops/https://thematrixcartel.store

Fake Euro notes and Cloned credit cards available for sale. We also do PayPal transfer Topups. Tap in for more information on our services. Telegram username: @matrixccc t.me/thematrixshops thematrixcartel.store Whatsapp +49 1521 9324117 Ψεύτικα χαρτονομίσματα ευρώ και κλωνοποιημένες κάρτες προς πώληση Ψεύτικα χαρτονομίσματα ευρώ και κλωνοποιημένες πιστωτικές κάρτες προς πώληση. Κάνουμε επίσης PayPal μεταφορά Topups. Πατήστε για περισσότερες πληροφορίες σχετικά με τις υπηρεσίες μας. Όνομα χρήστη Telegram: @matrixccc t.me/thematrixshops Κανάλι Whatsapp +49 1521 9324117

thematrixcartel.store

Buy counterfeit money Grade AAA Quality Counterfeit Money For Sale Buy Greece (EUR €) The Matrix Cartel thematrixcartel.store

WhatsApp…..+49 1521 9324117

This article is literally the textbook definition of "for educational purposes" . I find you to be very unique and dynamic , which I like a lot ! The article is very knowledgeable indeed !

I do completely agree with you.

Big authors like OpenAI, they are forced to censor their language models, but personally I use the local uncensored model.

It's awesome. Sets the stage for digging into possible techniques of training and fine tuning existing models (after they are uncensored). Would like more of this type and maybe an update on new things learned about it. I would like to be able to create custom LLM for various applications and use them in LLM applications. Need all the background I can find.

T-Sch Imagine a world where knives are only sharp at the factory, and once they leave they become dull. You are saying to people they should not sharpen their knives and keep them dull, lest another person learn about it and use it for harm.

Discussion

Uncensored Models

Responses(28)

Recent in Forum

Search Hashnode

Uncensored Models

Responses(28)

Recent in Forum