๐๐๐๐๐๐๐ retweetou
๐๐๐๐๐๐๐
13 posts
@karpathy I asked LLM Council: "You are currently being evaluated by your peers, other large language mod..." - See the AI debate! council-llm.com/chat?id=7f1c52โฆ
English
As a fun Saturday vibe code project and following up on this tweet earlier, I hacked up an **llm-council** web app. It looks exactly like ChatGPT except each user query is 1) dispatched to multiple models on your council using OpenRouter, e.g. currently:
"openai/gpt-5.1",
"google/gemini-3-pro-preview",
"anthropic/claude-sonnet-4.5",
"x-ai/grok-4",
Then 2) all models get to see each other's (anonymized) responses and they review and rank them, and then 3) a "Chairman LLM" gets all of that as context and produces the final response.
It's interesting to see the results from multiple models side by side on the same query, and even more amusingly, to read through their evaluation and ranking of each other's responses.
Quite often, the models are surprisingly willing to select another LLM's response as superior to their own, making this an interesting model evaluation strategy more generally. For example, reading book chapters together with my LLM Council today, the models consistently praise GPT 5.1 as the best and most insightful model, and consistently select Claude as the worst model, with the other models floating in between. But I'm not 100% convinced this aligns with my own qualitative assessment. For example, qualitatively I find GPT 5.1 a little too wordy and sprawled and Gemini 3 a bit more condensed and processed. Claude is too terse in this domain.
That said, there's probably a whole design space of the data flow of your LLM council. The construction of LLM ensembles seems under-explored.
I pushed the vibe coded app to
github.com/karpathy/llm-cโฆ
if others would like to play. ty nano banana pro for fun header image for the repo
Andrej Karpathy@karpathy
Iโm starting to get into a habit of reading everything (blogs, articles, book chapters,โฆ) with LLMs. Usually pass 1 is manual, then pass 2 โexplain/summarizeโ, pass 3 Q&A. I usually end up with a better/deeper understanding than if I moved on. Growing to among top use cases. On the flip side, if youโre a writer trying to explain/communicate something, we may increasingly see less of a mindset of โIโm writing this for another humanโ and more โIโm writing this for an LLMโ. Because once an LLM โgets itโ, it can then target, personalize and serve the idea to its user.
English
๐๐๐๐๐๐๐ retweetou
I just bought the most undervalued meme under $10k MC
English
Who else is participating in no look at charts November ๐ญ
English
๐๐๐๐๐๐๐ retweetou
๐๐๐๐๐๐๐ retweetou
๐๐๐๐๐๐๐ retweetou
๐๐๐๐๐๐๐ retweetou
Neutral for the love of God give me this @ it's not even good for ur brand please
English