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

Chain of Recursive Thoughts: Make AI think harder by making it argue with itself

github.com

Post reply on HN
Print view View original

231–240 of 253 posts

Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself

#231

Post by small_scombrus »

hu3 wrote: post #115

Here's some related challenge I'm facing. Maybe someone can help me: I also managed to make AI critique itself and that improved code generation a ton. For a TypeScript backend project that runs with Bun, I tell AI to also generate and run unit tests after every code change suggested by AI. How do you solve the risk of AI writting and executing unit tests with something like `rm -rf /` and wiping your files? Docker w…

> How do you solve the risk of AI writting and executing unit tests with something like `rm -rf /` and wiping your files?

The same way you stop any person or program or third party from doing something dumb or nefarious with your files. Don't give them any access to important files.

Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself

#232

Post by odie88 »

zoogeny wrote: post #152
odo1242 wrote: post #88

Something I do sometimes is: - Have an AI chat model come up with an answer to a problem. - Have it write a report discussing the details of the problem and why it's answer is correct, directed at a person or AI model who has no knowledge of the initial problem or technical field. - Have a second AI model with no knowledge of the problem grade the report, and write it's own report either (a) asking for clarification…

I do the same, and I have one other technique. I will often have a few chats going for a project, but with different contexts. For example, one might be tech focused, another marketing focused, another with some context on my personal goals, etc. So I will take the same question and feed it into the chats with differing context. It is almost like having different perspectives on the same problem. And the conclusions…

This is how I’ve been using Gemini and it’s the first time I’m really seeing consistent value.

I’ll get a context into a solid place with as much information as I can about a project. Usually getting up to 100k tokens.

Then I ask it to give me a summary I can use in a fresh chat, that will maintain the current context. This lets me reclaim space, bring responsiveness back to sane levels, have a baseline chat I use to spin up branches for marketing, design (it’s pretty helpful at trouble shooting Substance Designer graphs), etc.

I’ve found myself going into sub branches from there… like a marketing context that pushes branches into different marketing channels.

Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself

#233

Post by consumer451 »

voidspark wrote: post #221
casenmgreen wrote: post #210

Earlier quoted context omitted.

I may be wrong, but it seems to me this also is a case of improper use of words. Those LLMs neither agree nor disagree. They do not understand. They produce output , and we read that output and we ourselves consider the output to be something, or something else. All an LLM does is produce output. There's no conceptual understanding behind it, and so there is no agreement, or disagreement.

LLMs learn high-dimensional representations that capture conceptual relationships in their training data. They manipulate those representations in ways that approximate human reasoning.

> They manipulate those representations in ways that approximate human reasoning.

Fwiw, this is the story of my life. Seriously.

Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself

#234

Post by kmacdough »

senordevnyc wrote: post #183
moribunda wrote: post #135

Earlier quoted context omitted.

Exactly... I thought that implementing STORM was just a basic step in this topic... Looks like we're running in circles.

Mind sharing a link?

Here's a paper on agent architectures including multi agent. A bit old at this point, but a good overview.

https://arxiv.org/abs/2404.11584

Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself

#235

Post by autokad »

dudeinhawaii wrote: post #190

I see a lot of threads pitting models against each other (or whole swarms of them) in the hope that "wisdom of crowds" will magically appear. After a stack of experiments of my own—and after watching the recent ASU/Microsoft-Research work [1].. I've landed on a simpler takeaway: An LLM is a terrible verifier of another LLM. Subbarao Kambhampati's "(How) Do LLMs Reason/Plan?" talk shows GPT-4 confidently producing pro…

actually, I found that you can definitely yield better results. I ran an experiment with 1 prompt at temperature 0 and 9 with temperature 1.

I found the most anomalous response was as good (15/20) or better (5/20) than the temperature 0 response in 20 samples.

Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself

#236

Post by aoeusnth1 »

zurfer wrote: post #212
dudeinhawaii wrote: post #190

I see a lot of threads pitting models against each other (or whole swarms of them) in the hope that "wisdom of crowds" will magically appear. After a stack of experiments of my own—and after watching the recent ASU/Microsoft-Research work [1].. I've landed on a simpler takeaway: An LLM is a terrible verifier of another LLM. Subbarao Kambhampati's "(How) Do LLMs Reason/Plan?" talk shows GPT-4 confidently producing pro…

Your references show me that it is absolutely task depended. In many domains it's true that "criticizing is easier than creating". The best example might be books and movies, where it's trivial to say the characters were shallow, but it's surprisingly hard to create deeply interesting characters. In Software Engineering, there are similar dynamics. An LLM with a security vuln finding prompt will be able to point out…

In principle, Math proofs are another relatively easy to verify problem. In the extreme case, you can express any math proof as a computer-verifiable formalism — no intelligence necessary. Step back one step, and you could have a relatively weak model translate a proof into verifiable formalism and then use a tool call to run the verification. Coming up with the proof is an expensive search process, while verifying it is more mechanical. Even if it is not completely trivial to make the proof computer-verifiable, it might still be a vastly easier task compared to finding the proof in the first place.

Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself

#237

Post by corvus-cornix »

mikepurvis wrote: post #70
Lerc wrote: post #54

I kind of want to try something like this at a larger scale in an always-on mode where I have a 'senate' of debate. Rather than responding to prompts on a case by case basis, provide a list of tasks (potentially with deadlines) and let the senate work on them, break off into groups to manage subtasks, challenge results , make suggestions. Even potentially a tree of analysts where suggestions only gets passed up the t…

In doing some DevOps-y type tasks recently (ansible, packer, docker, baking images with guestfish), I've found it very frustrating how much ChatGPT will confidently tell me to use flags on tools that don't exist, or hallicinate completely non-existent functions or behaviours. And then when I spend time trying what it suggests only to hit a wall and come back like wtf mate it breezily goes "oh yes so you're right, goo…

I've also found LLMs to perform poorly at DevOps tasks. Perhaps there's a lack of training data. On the bright side this hints at better job security for platform engineers.

Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself

#238

Post by faramarz »

That's cool! thanks for making it easy to fork and play with this!

I've just begun my own iteration of adding Nash Equilibrium (NECoRT?) and reframing the "prompt engineering" to be a multi-agent negotiation. Curious what others think? https://github.com/faramarz/NECoRT/

my reasoning is that enterprise LLMs wont have any issue with the extra compute costs and would rather reconcile complex financials with various modeling optimizations.

I'm very new to public repo and contributions, and hope someone can point out if I'm doing it wrong.

my intention was to fork the ops codebase so I can test out my theory, and push as PR eventually

Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself

#239

Post by the_af »

visarga wrote: post #201
the_af wrote: post #127

Earlier quoted context omitted.

Does it share context even if no "memory updated" message appears indicating it has stored a fact about you? I asked ChatGPT and it says no, but then again it's not reliable at introspection or at revealing data about how it works.

I think they are different systems, one is a collection of saved snippets and the other more like RAG over chat history.

ChatGPT assures me it doesn't use RAG (fed from my other chat windows), but will use memory-saved preferences (in the store that can be accessed and reviewed in Settings->Personalization->Memory).

Then again, I don't think ChatGPT is reliable when reporting on its own inner workings.

---

Oh, no, here it says it also references chat history: https://help.openai.com/en/articles/8590148-memory-faq

Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself

#240

Post by foobiekr »

dudeinhawaii wrote: post #190

I see a lot of threads pitting models against each other (or whole swarms of them) in the hope that "wisdom of crowds" will magically appear. After a stack of experiments of my own—and after watching the recent ASU/Microsoft-Research work [1].. I've landed on a simpler takeaway: An LLM is a terrible verifier of another LLM. Subbarao Kambhampati's "(How) Do LLMs Reason/Plan?" talk shows GPT-4 confidently producing pro…

"letting GPT-4 critique its own answers reduces accuracy"

This is because the output, being the input, steers directly into the tree as soon as the tree is in the context window.

Post reply on HN