[flagged]
Or "thinking" just got a new meaning and it's to convey information in the field - perhaps the Oxford dictionary will add it soon?
Chain of Recursive Thoughts: Make AI think harder by making it argue with itself
21–30 of 253 posts
Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself
#22This seems to be different than I expected from the title. I thought it would be explicitly adversarial.
1. You are the assistant. Please answer the question directly.
2. You are the cross-examiner. The assistant is wrong. Explain why.
3. You are the assistant. The cross-examiner is wrong. Defend your claim.
4. You are a judge. Did either party make their case, or is another round of argumentation required?
I haven't tried this. No idea if it works. But I find it's helpful to ask ChatGPT, in separate prompts, "XYZ is true, explain why" and "XYZ is false, explain why" and see which one seems more convincing.
Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself
#23So glad to see a write up on this finally. I'm no machine learning phd but I always wondered why this wasn't more of a thing. Like an extension of a GAN conceptually, sort of, not really at all Im sure.
Also I think I kind of assumed OpenAI might be doing this behind the curtain?
Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself
#24I’ll second this. I often use a “research assistant “ and skeptical“department head” personas working together/against each other as a research team. It works well and is occasionally hilarious, replete with the occasional HR complaint when things go off the rails. ( I typically use local uncensored models)
Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself
#25[flagged]
Why is it so hard to believe that a complex neural network can think? You literally have one over your shoulders that does exactly that.
Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself
#26this is amazing - I love seeing novel approaches to optimizing
Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself
#27[flagged]
Don't people get tired of having this same "debate" on every post about LLMs? And I scare quote debate because the naysayers never support their strong claims beyond the most superficial of responses. It's all just so tiring at this point.
Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself
#28I think this is how we get ML models to come up with novel ideas. Diagonalize against all the ideas they’ve already tried and dismissed via self-argument but keep certain consistency constraints. (Obviously much easier said than done.)
What you just said is what I tried and failed to say ten minutes ago! https://news.ycombinator.com/item?id=43835798
It’s working! Oh, wait …
These models have limitations obviously, but many critiques apply equally or more to people.
If people were tasked with one shot, 10 second answers, to be written out in near errorless grammar, the LLM’s viewing our responses to prompts would be spending a lot of time discussing our limitations and how to game us into better responses. Humor, not at all humor.
Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself
#29Earlier quoted context omitted.
_#_ has to be one of the worst word shortening schemes I've ever seen get widespread. It only works with a very small number of long-lived technologies, in which case they basically just get a nickname, "k8s" "i18n". It does not at all work for larger contexts. You're basically making someone solve a crossword (2 across, 10 letters with two filled in) just to parse your sentence.
I just googled it and it looks like “n8n” is the name of the service. The op wasn’t abbreviating anything so I don’t think it’s the same phenomenon as what you’re describing.
Well, the service is doing the same thing though. The part I don't understand is that I assume n8n is short for "Nation" but literally every single person I've seen talk about it on YouTube (which is quite a lot) say "En Eight En" every time.
Re: Chain of Recursive Thoughts: Make AI think harder by making it argue with itself
#30I'm having a lot of fun experimenting with stuff like this. I'm trying to put together an unrealengine blueprints style graph editor to allow people to design workflows like this where you start with the user prompt input, which goes to one agent, which makes an initial attempt, and then that conversation history gets passed to another "agent" with a different system prompt telling it to be a harsh critic, but to als…
I am thinking the same thing! Multiple "personalities", in parallel, or in series. For example, I have approximated, in GPT, some of Gemini's ability to call out nonsense, sloppy thinking, by telling GPT to be mean! (The politeness seems to filter out much that is of great value!) However, the result is not pleasant to read. Gemini solved this in their training, by doing it in two phases... and making the first phase…
> As a side note, I think there would be great value in a UI that allows a "group chat" of different LLM personalities.
This is the basic idea behind autogen. They also have a web UI now in autogen studio, it's gotten a bit better. You can create "teams" of agents (with different prompts, themes, tools, etc.) and have them discuss / cooperate. I think they even added memory recently. Have a look at it, might be what you need.