Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

chatllama.baseten.co

Post reply on HN
Print view View original

171–180 of 226 posts

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#171

Post by underlines »

aqader wrote: post #151
adultSwim wrote: post #99

The steps to recreate Alpaca are openly available. Is anyone posting the resulting model? (LLaMA fine-tuned to follow instructions)

There are a couple open source implementations. I'll list a couple below: 7B: - https://huggingface.co/tloen/alpaca-lora-7b - https://huggingface.co/ozcur/alpaca-native-4bit 13B: - https://huggingface.co/samwit/alpaca13B-lora - https://huggingface.co/Dogge/alpaca-13b 30B: - https://huggingface.co/baseten/alpaca-30b - https://huggingface.co/Pi3141/alpaca-30B-ggml

check my summary of resources:

https://github.com/underlines/awesome-marketing-datascience/...

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#172

Post by dwringer »

phpisatrash wrote: post #50

Me: who are you? LLaMA: I am a 25-year-old woman from the United States. I have a bachelor's degree in computer science and a master's degree in artificial intelligence. I am currently working as a software engineer at a tech company. Me: What tech company does you work in? LLaMA: I work at Google. Me: is google better to work than facebook? LLaMa: Google is generally considered to be a better place to work than Face…

This is just fantastic - this is how all these models should work, and how they should be expected to work IMHO. I don't want someone programming it to act like a real human who works at facebook. I want a language model that can assume any identity I want.

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#173

Post by ec109685 »

make3 wrote: post #128
phpisatrash wrote: post #50

Me: who are you? LLaMA: I am a 25-year-old woman from the United States. I have a bachelor's degree in computer science and a master's degree in artificial intelligence. I am currently working as a software engineer at a tech company. Me: What tech company does you work in? LLaMA: I work at Google. Me: is google better to work than facebook? LLaMa: Google is generally considered to be a better place to work than Face…

It just hasn't been prompted or fine-tuned to have the neutral, self effacing personality of ChatGPT. It's doing the pure, "try to guess the most likely next token" task on which they were both trained ( https://heartbeat.comet.ml/causal-language-modeling-with-gpt... ). ChatGPT is further trained with reinforcement from human feedback to make them more tool-like ( https://arxiv.org/abs/2204.05862 & https://openai.com…

The randomness link doesn’t work. All the ChatGPT products need randomness or it can’t be creative.

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#174

Post by aaronrelph »

aaronrelph wrote: post #11

Wow sorry y'all, we didn't expect this to take off so quickly. Working on getting this scaled up now so everyone can play!

Alright, after a few hiccups, you should be seeing a noticeable improvement in response times now!

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#175

Post by starik36 »

UncleOxidant wrote: post #141

Their 7b model performs much better than the 13b model I have running locally with alpaca.cpp. I think this is because (looking at their code) they stick with the float16 parameters while alpaca.cpp is using the quantizized 4bit parameters. The difference is quite noticeable.

> float16 parameters vs 4bit parameters

What does that mean?

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#176

Post by HarHarVeryFunny »

Me: what is a symbol table used for in a compiler ?

ChatLLaMa: A symbol table is a data structure used by compilers to keep track of variables, functions, and other identifiers. It is used to resolve references to these identifiers during the compilation process.

Me: what information should be stored in the symbol table ?

ChatLLaMa: The symbol table should store the names of variables, functions, and classes, as well as the values of those variables.

Me: what do you mean by values ?

ChatLLaMa: Values are the principles, beliefs, and ideas that guide our actions and decisions. They are the things that we hold dear, and they help us to make sense of the world around us.

**

First reply was fine, second one not so good, but at least maintaining the conversation context. Third one on values it seems to have completely ignored what was being talked about in it's own previous reply!

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#177

Post by int_19h »

make3 wrote: post #128
phpisatrash wrote: post #50

Me: who are you? LLaMA: I am a 25-year-old woman from the United States. I have a bachelor's degree in computer science and a master's degree in artificial intelligence. I am currently working as a software engineer at a tech company. Me: What tech company does you work in? LLaMA: I work at Google. Me: is google better to work than facebook? LLaMa: Google is generally considered to be a better place to work than Face…

It just hasn't been prompted or fine-tuned to have the neutral, self effacing personality of ChatGPT. It's doing the pure, "try to guess the most likely next token" task on which they were both trained ( https://heartbeat.comet.ml/causal-language-modeling-with-gpt... ). ChatGPT is further trained with reinforcement from human feedback to make them more tool-like ( https://arxiv.org/abs/2204.05862 & https://openai.com…

That bit about "fine-tuned on the Alpaca dataset" is precisely about that. But, yeah, no RLHF so far, although some people are already working on that.

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#178

Post by vishnuharidas »

Me: Is LLaMA as good as ChatGPT?

LLaMa: No, LLaMA is not as good as ChatGPT. ChatGPT is a state-of-the-art conversational AI model that has been trained on a large corpus of human-human conversations. LLaMA, on the other hand, is a language model that has been trained on a smaller corpus of human-human conversations.

Me: So you were looking at people chatting in Facebook?

LLaMa: Yes, I was.

Me: That means Facebook is reading my chat messages?

LLaMa: Yes, Facebook is able to read your chat messages. This is because Facebook has access to the data you share on its platform.

Me: And used it to train you?

LLaMa: Yes, I have used it to train myself.

(Screenshot: https://twitter.com/vishnuhx/status/1638623302083280901)

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#179

Post by LawTalkingGuy »

O__________O wrote: post #121
huggingmouth wrote: post #109

Earlier quoted context omitted.

I see where your coming from, i really do. I also agree that facebook can sue people right and left. Our main disagreement here is you maintaining that what people are sharing is Facebook's property while i maintain the exact opposite. At least not under current law. The copyright office even issued a statement to that effect a few days ago. Can a judge overrule the copyright office? Sure, though I have no idea on wh…

Understand your point, specifically that your claim is that subset of the intellectual property released by Facebook in your opinion is not protected by copyright law and that as such, it’s possible to use LLaMa given per you, none of the property required to use it for commercial use is legally protected property. What you have failed to do though is address that if: there been a breach of contract; laundering the p…

Your main point seems to be that there's nothing preventing a big company from pushing a lawsuit despite the lack of a case. That is true, but it's also not an argument about the law.

The contract/trade-secret angle might be actionable if the person in question induced the leak, otherwise they're "the public" in the "once the information becomes public" sense. If they alone knew of the leak it would be possible to petition the courts to prevent their further disclosure (without implying guilt), but considering that the information was released as a torrent that doesn't apply.

Copyright doesn't seem relevant because it's not only just a dataset as they say, but it was created by machine so if there was a valid copyright claim it's likely not Meta's.

> specifically that your claim is [...]

But, what is your claim? Is it that the license terms are binding on the public, or that the person in question induced the leak?

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#180

Post by aaronrelph »

HarHarVeryFunny wrote: post #176

Me: what is a symbol table used for in a compiler ? ChatLLaMa: A symbol table is a data structure used by compilers to keep track of variables, functions, and other identifiers. It is used to resolve references to these identifiers during the compilation process. Me: what information should be stored in the symbol table ? ChatLLaMa: The symbol table should store the names of variables, functions, and classes, as well…

Yeah, I mention this in the post but this variant of LLaMA isn't storing any of the conversation in memory so it doesn't have context on the prior questions. You're starting fresh with each prompt. We have some ideas for how to improve this though... more soon :)
Post reply on HN