g1: Using Llama-3.1 70B on Groq to create o1-like reasoning chains
github.com
github.com
TreeOfThoughts is a more sophisticated method, see - https://arxiv.org/pdf/2305.10601
The clue we all had with OpenAI for a long time that this was a search through a tree, they hired Noam Brown, and his past work all hinted towards that. Q, is obviously a search on a tree like A. So take something like CoT, build out a tree, search for the best solution across it. The search is the "system-2 reasoning"
The chain of thought would be the final path through the tree. Interactively showing the thought tokens would give the game away, which is why they don’t show that.
In this case, the model would explore several chain of thoughts during training, but only output a single chain during inference (as the sibling comment suggests).
We can say the same thing about RL implying PPO, however there’s pretty big hints, namely Noam Brown being involved. Many of the things Noam Brown has worked on involve RL in tree search contexts.
He has also been consistently advocating the use of additional test-time compute to solve search problems. This is also consistent with the messaging regarding the reasoning tokens. There is likely some learned tree search algorithm, such as a learned policy/value function as in AlphaGo.
It’s all speculation until we have an actual paper. So we can’t categorically say MCTS/learned tree search isn’t involved.
Did they maybe expand to a tree during training to learn more robust reasoning? Maybe. But it still comes down to a regular transformer model at inference time.
I've gotten mini to think harder by asking it to, but it didn't make a better answer. Though now I've run out of usage limits for both of them so can't try any more…
nobody except for recent deepmind research has shown test time scaling like o1
https://arxiv.org/pdf/2403.09629
> In the Self-Taught Reasoner (STaR, Zelikman et al. 2022), useful thinking is learned by inferring rationales from few-shot examples in question-answering and learning from those that lead to a correct answer. This is a highly constrained setting – ideally, a language model could instead learn to infer unstated rationales in arbitrary text. We present Quiet-STaR, a generalization of STaR in which LMs learn to generate rationales at each token to explain future text, improving their predictions.
>[...]
>We generate thoughts, in parallel, following all tokens in the text (think). The model produces a mixture of its next-token predictions with and without a thought (talk). We apply REINFORCE, as in STaR, to increase the likelihood of thoughts that help the model predict future text while discarding thoughts that make the future text less likely (learn).
e: answer to my own question https://x.com/_xjdr/status/1835352391648158189
You don't see the scaling with respect to token length with non-FT'd CoT like this, in my opinion.
I'm just saying whatever they did in their [new] model, I think they also added CoT on top of it, as the outer layer of the onion so to speak.
You will not unlock "o1-like" reasoning by making a model think step by step. This is an old trick that people were using on GPT3 in 2020. If it were that simple, it wouldn't have taken OpenAI so long to release it.
Additionally, some of the prompt seems counterproductive:
>Be aware of your limitations as an llm and what you can and cannot do.
The LLM doesn't have a good idea of its limitations (any more than humans do). I expect this will create false refusals, as the model becomes overcautious.
Can it not be trained to do so? From my anecdotal observations, the knowledge cutoff is one thing that LLMs are really well trained to know about. Those are limitations that LLMs are currently well trained to handle. Why can it not be trained to know that it is quite frequently bad at math, it may produce sometimes inaccurate code etc.
For humans also, some people know some things are just not their cup of tea. Sure there are times people may have half baked knowledge about things but one can tell if they are good at XYZ things, and not so much at other things.
A LLM has a huge amount of data ingested. It can create character profiles, audience, personas etc.
Why wouldn't it have potentially even learned to 'understand' what 'being aware of your limitations' means?
Right now for me 'change of reasoning' feels a little bit of quering the existing meta space through the reasoning process to adjust weights. Basically priming the model.
I would also not just call it a 'trick'. This looks simple, weird or whatnot but i do believe that this is part of AI thinking process research.
Its a good question though what did they train? New Architecture? More parameters? Is this training a mix of experiments they did? Some auto optimization mechanism?
But if you read all the knowledge of humans, were does your reasoning start? Probably at a very high level of it.
If you look at human brains, we conduct experiments right? As a software developer, we write tests. ChatGPT can already run python code and it can write unit tests.
We do not use proofs when we develop. An AI could actually doing this. But at the end its more of a question who does it better, faster and cheaper eh?
Humans do in most cases have some knowledge about why they know the things they know. They can recall the topics they learned at school, and can deduce that they probably heard a given story from a friend who likes to discuss similar topics, etc.
LLMs have no access to the information they were trained on. They could know that everything they know was learned during the training, but they have no way of determining what they learned about and what they didn't.
It's only when we can interact with the environment to test our hypothesis that we then refine what we know and update our priors appropriately.
If we let LLMs do that as well, by allowing it to run code and interact with documentation/the internet and double-check things its not sure of, it's not out of the question LLMs won't eventually be able to more reliably understand its limitations.
Humans usually know (at least roughly) the source of anything they know, as there will be a memory or a known event associated with that knowledge.
LLMs have no analogous way to determine the source of their knowledge. They might know that all their knowledge comes from their training, but it has no way of knowing what was included in the training and what wasn't.
This could maybe be achieved with some more fancy RAG systems, or online training abilities. I think an essential piece is the ability to know the source of information. When LLMs reliably do, and apply that knowledge, they'll be much more useful. Hopefully somebody can achieve this.
alignment is a tough problem and aligning long reasoning sequences to correct answer is also a tough problem. collecting high quality CoT from experts is another tough problem. they started this project in october, more than plausible it could take this time
Or it's a unique advantage because this stuff doesn't happen without good researches who may want:
1) Their name in scientific papers
2) They might actually care about the openess of AI
Playing catch up and trying to attract talent from the hot-new-thing OpenAI requires incentives beyond lots of money. I contend actually being open helps.
I'm sure that's one reason Facebook has an open source model, scientists can care about ethics and could be attracted to openness.
The "Attention Is All You Need" guys all worked at Google. Google is where they are despite having the smart guys with a new idea a few years back.
Of course, IMHO it wouldn't have have helped Google if they'd kept the transformer architecture secret. They'd have fumbled it because they didn't realise what they had.
Only then did Google realise there might be something to this LLM stuff - so they responded with Bard, a product so poorly received they later had to completely rebrand it. Looks like they didn't have a "sentient" model up their sleeve after all. Then the updated, rebranded model had a bunch of image generation embarrassments of its own.
Admittedly, they have recovered somewhat since then; they're second on some performance leaderboards, which is respectable.
But there was a real tortoise-and-hare situation where they thought they were so far ahead they had time for a nap, until they got overtaken. Any lead they had from inventing transformers and being the only people with TPUs has been squandered.
This kind of thing is still so funny to me.
I wonder if the first guy who gets AGI to work will do it by realizing that he can improve LLM reliability over some threshold by telling it in all caps that his pet's life depends on the answer.
(But it all depends on the fine-tuning they did, so who knows, maybe it's just an Easter egg)
> In order to make the draft response nicer and complete, a set of question [sic] and its answer are provided," reads one prompt. "Please write a concise and natural reply by modify [sic] the draft response," it continues.
This really sounds like a placeholder made up by one engineer until a more qualified team sits down and defines it.
The butterfly keyboards were unusable to me and also the OS got too locked down so I left the platform.
Sort of like how offering to pay the LLM $5 improves its output. The LLM's taking your prompt seriously, but not literally.
Ask an LLM what hallucination is, ask it to write a story with etc.
without zeroing out things, everything has and can have some impact
If so, I imagine o1 clones could just be fine tunes of llamas initially.
Example prompt for that: "give me three sentences that end in 'is'."
Does Midjourney output look like an average human drawing?
Obviously, OpenAI knows how to train a classifier...
No, perhaps because it's heavily trained on photos.
A LLM can therefore have an higher IQ because it can combine all fields.
Also parameters and architecture might or might not be a limiting factor to us humans or a LLM. But LLM and parameter size, optimizations etc. are just at the beginning.
If we now have a good reasoning llm, we can build more test data automatically. Basically using the original content + creating new ones which can then lead to new knowledge = research.
It might be the 200M user base of OpenAI that provided the necessary guidance for advanced CoT, implicitly. Every user chat session is also an opportunity for the model to get feedback and elicit experience from the user.
I'm not convinced OpenAI is using one model. Look at the thinking process (UI), which takes time, and then suddenly, you have the output streamed out at high speed.
But even so, people are after results, not really the underlying technology. There is no difference of doing it with one model vs multiple models.
According to OpenAI, the model does it's thinking behind the scenes, then at the end summarizes that thinking for the user. We don't get to see the original chain-of-thought reasoning, just the AI's own summary of that reasoning. That explains the output timing.
If they were just doing prompt engineering and multiple inferences they'd definitely want to keep that a competitive secret and send all the open source devs off in random directions, or keep them guessing, rather than telling them which way to go to replicate Q-Star.
This is still clearly CoT, with all its limitations and caveats as expected. That's an improvement, sure, but definitely not a qualitative leap like OAI is trying to present it. (in a really shady manner)
Saying it's just CoT is kind of meaningless. Even just looking at the examples on Open AI's blog and you quickly see no other model today can generate or utilize CoT to anywhere near that quality through prompting or naive fine-tuning.
Starting from the strawberry example: it counts 3 "r"s in "strawbery", because the training makes it ignore grammatical errors if they're not relevant to the conversation (which makes sense in an instruction-tuned model) and their CoT doesn't catch it because it's not specialized enough. Will this scale with more compute thrown at it? I'm not sure I believe their scaling numbers. The proof should be in the pudding.
I've also had mixed results with coding in Python, it's barely better than Sonnet in my experience, but wastes a lot more tokens, which are a lot more expensive.
They might have improved things and made a SotA CoT that works in tandem with their training method, but that is definitely not what they were originally hyping (some architecture-level mechanism, or at least something qualitatively different). It also pretty obviously still has limited compute time per token and has to fit into the context (which is also still suffering from the lost-in-the-middle issue by the way). This puts the hard limit on the expressivity and task complexity.
That's an incredibly cynical choice of phrasing.
Of course they don't want to help the competition, that's what a competition is. The competition isn't helping OpenAI either.
All the problems with llm not reasoning (like planning, counting letters or deductive inference) are easy for classical algos. There needs to be a way to split the thinking process into two parts and then execute each part on the appropriate model.
Two centuries ago there were no computers, everything had to be done by humans. Get to that level first before you whip out code.
Not updated the Readme yet
Thanks for the fork and the suggestions though - looks like I'll be having fun with this over the week!
It’s a way to convert a text or response into an array of numbers, that can be used for similarity lookups.
I made a way to query large datasets of text strings: https://github.com/punnerud/search-embeddings-llama3.1
Can be used to let it explore a graph of knowledge as long as the graph is related to the original question, and can explore different solutions at the same time without repeating itself (then it’s get linked back to similar answers and stopped)
> Result: .9 is larger than .11
we've broken the semver barrier!
I think this class of problem might be better solved by allowing the LLM to 'zoom in' and view the input differently. Rather like you might peer closer for more detail if someone asked you about the print quality of something you were reading.
'zoom in' could input the same text letter by letter, or even in image form (rasterize the text) to help answer questions like "How many letters in the word strawberry contain straight lines?"
The idea is not silly in my view, I did something similar here: https://github.com/pseudotensor/open-strawberry
The idea is that data generation is required first, to make the reasoning traces. ToT etc. are not required.
The core innovation [1] of o1 lies in its ability to generate and refine internal chains of thought before producing a final output [2]. Unlike traditional LLMs that primarily focus on next-token prediction, o1 learns to:
1. Recognize and correct mistakes 2. Break down complex steps into simpler ones 3. Try alternative approaches when initial strategies fail
This process allows o1 to tackle more complex, multi-step problems, particularly in STEM fields.
OpenAI reports observing new "scaling laws" with o1 [5]:
1. Train-time compute: Performance improves with more extensive reinforcement learning during training. 2. Test-time compute: Accuracy increases when the model is allowed more time to "think" during inference.
This suggests a trade-off between inference speed and accuracy.
Sources [1] Introducing OpenAI o1 https://medium.com/%40sriramramakrishnan.aiexpert/openais-o1... [2] Learning to Reason with LLMs | OpenAI https://openai.com/index/learning-to-reason-with-llms/ [3] OpenAI o1 models - FAQ [ChatGPT Enterprise and Edu] https://help.openai.com/en/articles/9855712-openai-o1-models... [4] OpenAI releases new o1 reasoning model - The Verge https://www.theverge.com/2024/9/12/24242439/openai-o1-model-... [5] 9 things you need to know about OpenAI's powerful new AI model o1 https://fortune.com/2024/09/13/openai-o1-strawberry-model-9-... [6] Notes on OpenAI's new o1 chain-of-thought models https://simonwillison.net/2024/Sep/12/openai-o1/ [7] OpenAI just dropped o1 Model that can 'reason' through complex ... https://www.tomsguide.com/ai/openais-o1-model-takes-ai-to-a-... [8] Models - OpenAI API https://platform.openai.com/docs/models [9] OpenAI Unveils O1 - 10 Key Facts About Its Advanced AI Models https://www.forbes.com/sites/janakirammsv/2024/09/13/openai-...
o1 is far more than just CoT mechanics. It relies on a specialized model or collection of models that offer new capabilities to make CoT work far better than it works with a stock LLM.
For instance, o1 can recognize and correct its own mistakes and it seems to know how to dig deeper when needed. That's not something that stock LLMs do very well.
also I think they deliberate give you bad answers sometimes / a lot over the last year to build up advanced chains where the user is not getting what they want so you have to explain why. I started building up like 10 or so of these conversations where after like 100 messages it gets the right answer and it was like hmm, I wonder if they are using this.
just my rambles
Did you benchmark your system against MMLU-pro?
Because it says so nowhere in the repo.
Man Elon makes things confusing.
Tailoring prompts is likely still the best way to maximize performance when you can, but in broader domains you'd work around this through strategies like asking the LLM to combine predefined reasoning modules, or creating multiple reasoning chains and merging/comparing them, explicit MCTS etc. I think those strategies will still be useful for a good while, but pieces of that search process, especially directing the search more efficiently, move to the LLMs over time as they get trained with this kind of data.
They didn't train a model for millions from experts to just basically use CoT now. Thats a harsh simplification, probably.
You are an expert AI assistant that explains your reasoning step by step. For each step, provide a title that describes what you're doing in that step, along with the content. Decide if you need another step or if you're ready to give the final answer. Respond in JSON format with 'title', 'content', and 'next_action' (either 'continue' or 'final_answer') keys. USE AS MANY REASONING STEPS AS POSSIBLE. AT LEAST 3. BE AWARE OF YOUR LIMITATIONS AS AN LLM AND WHAT YOU CAN AND CANNOT DO. IN YOUR REASONING, INCLUDE EXPLORATION OF ALTERNATIVE ANSWERS. CONSIDER YOU MAY BE WRONG, AND IF YOU ARE WRONG IN YOUR REASONING, WHERE IT WOULD BE. FULLY TEST ALL OTHER POSSIBILITIES. YOU CAN BE WRONG. WHEN YOU SAY YOU ARE RE-EXAMINING, ACTUALLY RE-EXAMINE, AND USE ANOTHER APPROACH TO DO SO. DO NOT JUST SAY YOU ARE RE-EXAMINING. USE AT LEAST 3 METHODS TO DERIVE THE ANSWER. USE BEST PRACTICES.
Does it work? Well not really:
https://lluminous.chat/?sl=Yjkxpu
https://lluminous.chat/?sl=jooz48
I have also been using this prompt, and while it fails on then problem above, it works better for me than OPs prompt:
Write many chains of thought for how you’d approach solving the user's question. In this scenario, more is more. You need to type out as many thoughts as possible, placing all your thoughts inside <thinking> tags.
Your thoughts are only visible to yourself, the user does not see them and they should not be considered to be part of the final response.
Consider every possible angle, recheck your work at every step, and backtrack if needed.
Remember, there are no limits in terms of how long you can think - more thinking will always lead to a better solution.
You should use your thoughts as a scratchpad, much like humans do when performing complicated math with paper and pen. Don't omit any calculation, write everything out explicitly.
When counting or maths is involved, write down an enormously verbose scratchpad containing the full calculation, count, or proof, making sure to LABEL every step of the calculation, and writing down the solution step by step.
Always remember that if you find yourself consistently getting stuck, taking a step back and reconsidering your approach is a good idea. If multiple solutions are plausible, explore each one individually, and provide multiple answers.
Always provide mathematical proofs of mathematical answers. Be as formal as possible and use LaTeX.
Don't be afraid to give obvious answers. At the very very end, after pages upon pages of deep thoughts, synthesize the final answer, inside <answer> tags.
* give me three sentences that end in "is"
* tell me the line of Star Spangled Banner that comes before "gave proof through the night"
But they did some good thinking before failing at it…
It's just a pile on of trial and error instructions (maybe learned from previous 'projects', but). There is no magic or skill to prompt 'engineering' anywhere.
Grok rhymes with cock, because Elon wants you to use it with your cock out.
That’s how I remember the difference.
Does the LLM take advantage of this? I don't know. It wouldn't surprise me if it did, and if it doesn't now I'd bet it will in the future. Either way though, throwing away those other languages could make the model dumber. As you allude to, there's a balance between intelligence and knowledge.
(in case you hadn't thought of it, those 'tools' can also be other LLMs with more specialized knowledge in a particular field. For example a 'translator' model)
Other 'facts' could also have more merit than it would first appear. Sure, one particular person's shoe size might not be needed, but if you were to filter out shoe sizes in general then the model might not be able to suggest how to find properly fitting footwear, or might not suggest that your back pain could be related to your shoes.
> That would be a worthwile endeavor and maybe even possible without boiling the oceans.
I think it's important to keep in mind that we're very early in the AI journey. Look at the power requirements of early computers versus the ones we use today. I'm all for keeping energy usage in mind, but I'd be careful with hyperbolic language as things are changing so quickly. Tasks that would have taken multiple GPUs can now run on my laptop CPU.
[1] https://www.edge.org/conversation/lera_boroditsky-how-does-o...
> I think it's important to keep in mind that we're very early in the AI journey.
That's what I am saying. At the moment there is this one really dumb idea, that bigger is better.