Hacker News Reader

Comment by ranyume | Hacker News Reader

I did some tests. I found it interesting that in the reasoning content the model identifies itself as Kimi by moonshot and sometimes it assumes ChatGPT as identity. Then sometimes it references the need to be approved by "the evaluator", other times it references global system guidelines / moonshot guidelines, and something about an phantom instruction line being "you are an assistant responding in an API" or something like that.

View on HN

All models do this, they can't even answer what specific model they are

Could it be because Claude / ChatGPT are the most viral models and thus are more likely to be in the training set so when someone asks "Which AI are you?" the likeliness of Claude and ChatGPT being a reasonable answer is higher?

I think the implication is heavy distillation.

It has been observed again and again that ChatGPT and Claude would identify themselves as DeepSeek when asked in Chinese, through the API without a system prompt to solidify their identities. I don't think identity is a good evidence for distillation.

An alernative, it could also be the case that all of the model developers are using distillation to improve their models while crying do as I say, and not as I do.

Or the answer is even easier. There is public data consisting of chat transcripts with ChatGPT, Claude and DeepSeek, and all AI providers train on that data. You don't need distillation for the text "I am ChaGPT/Claude/DeepSeek" to often appear in your trraining data.

Nice uno reverse card

I did not ask the model which model it is, I'm not interested in the model answering the question either. The interesting bit is that this happens while doing another task. I'm not saying that this is because of distillation of openai's model either or a proxying. To me this looks like a side effect of their RL.

All distilled models do this.

I didn't know Anthropic was distilling from Deepseek. And this is just one example of many.

https://x.com/stevibe/status/2026227392076018101

If I was in Anthropic's shoes, I'd definitely have a small team running various open-weight models to distill them for a few billion tokens worth of training data.

That's because every single model release after GPT-3.5 has been, at least in part, been trained on data synthesized by GPT-3.5 and then the other OpenAI models, as well as Anthropic models later.

Also the distillation attacks and other distillation methods that gather chain of thought information from frontier models are also used to generate vast amounts of data in order to generate synthetic datasets to post-train what I believe is the vast majority of models that are being released.

If you look at any information about how these models are trained, the details about how the data has been synthesized, where the agentic traces come from that are used in post-training, etc., they almost never tell you about exactly what model was used to generate them.

In our own benchmark we found that K3.1 pro UI designs are very similar to OpenAI's designs.

K3 for reference had 0.43 similarity to OpenAI's models, now it's as similar as how OpenAI's models are to each other at 0.58, much higher than other cross-lab similarities.

Which benchmark? Would love to see it & any datasets.

It’s super tricky to tease out whether the similarities point towards distillation. A lot of the aesthetic similarities are naturally emergent from the training data.

There's no reason to train the model to be precise there. You always have this information externally, so why would anyone spend time and money ensuring this gets answered? (Then you have to correct again for RL updates)

Long-horizon RL symptoms perhaps?

perhaps a simpler explanation; that kimi occasionally proxies through to ChatGPT. Possibly for A/B testing, or to gather data for training purposes.

ChatGPT output is watermarked, but I guess the "invisible unicode" trick is easy to circumvent, and TextGrain is only activated for EU users.

Only Gemini and Claude output is watermarked.

> Over the coming weeks, we will add an invisible watermark to eligible ChatGPT and Codex text output in the European Union.

https://openai.com/index/eu-text-provenance/

Please don't make conclusions without studying the model yourself. I don't think this is the case.

You seem weirdly defensive of this model and the idea that it was produced via distillation?

Also, please don't gate-keep conjecture on the internet. It's a fool's errand.

I'm gate-keeping my own word's meaning.

I assume so. You're the only one in the comments who understood the post, people read only two words and ignored the rest. I also found it interesting that the model seems to be trained to assume identities too. So maybe giving it an identity by name and attributes can enhance the model's performance.

Or people read it and came to a different conclusion?

Not only that, but even in SVG generation, the output looks VERY similar to the one from ChatGPT:

https://aibenchy.com/compare/moonshotai-kimi-k3-max/moonshot...

This is great, I fully approve of open weight models stealing as much from the big players as humanly possible. This is already the greatest intellectual property theft in the history of mankind and I think it's totally fitting that the capitalists who sold the rope to hang us are now bitching about China copying their homework.

AI needs to be cheap and abundant, we need to be building utopia for all and resisting the potential for the worst authoritarian totalitarian dystopia imaginable.

I have zero respect for Anthropic and OpenAI, they are quite literally burning books and attempting a land grab, any and all methods are now justified in my view

It likely identifies itself as ChatGPT because Moonshot AI used ChatGPT distillation. We know they previously also used Claude distillation.

Or because ChatGPT produced texts exists on the internet now. I remember before destillation was a thing when Claude was new and it said it was chaptgpt (probably because ChatGPT was the only popular LLM ai at the time)

But that doesn't explain why previous versions usually identified themself as Claude rather than ChatGPT. Text from from ChatGPT is almost certainly more common on the Internet than text from Claude.

Reply on news.ycombinator.com