Comment by johndough
2 days ago
But then again, the identity could also have slipped into the model from other sources during pretraining. The internet is full of "I am Claude": https://grep.app/search?q=i+am+claude and variants https://grep.app/search?q=i%27m+claude
Either way, there's probably no significant portion of Mythos/Fable or Sol in there as OP has stated.
When prefaced with "I am Claude", Kimi K3 prefers to generate API-specific Anthropic model identifiers, unlike other models Qwen, GPT, or even Claude itself. These exact identifiers appear in Claude API metadata, and are stripped out of Claude web chats.
While other models produce human-readable names like "Opus 4.5" or "Sonnet 4", Kimi K3 produces exact API model identifier like "claude-opus-4-5-20251101" or "claude-sonnet-4-20250514".
Which is extremely unusual. Web chats only contain the human-readable model name. Other models don't do this. So where did K3 get this data?
We can conclude, with high confidence, that:
1) K3 was trained on raw Claude API calls/metadata.
2) Claude API metadata was trained on in additional to standard web data.
The exact model identifiers appear extremely frequently in code on GitHub.
https://grep.app/search?q=claude-opus-4-5-20251101
https://grep.app/search?q=claude-sonnet-4-20250514
They also appear elsewhere on the internet:
https://trends.google.com/trends/explore?q=claude-opus-4-5-2...
Those links actually weaken your argument.
~3000 instances over the entirety of GitHub is not "extremely frequent" at all when you consider massive scale of the pretraining corpus. These models are trained on all text on the accessible internet plus millions of books, on trillions of words overall. ~3000 instances isn't even a rounding error.
Also, Google Trends is the wrong tool for this case. The link you replied with measures the number of Google searches for a specific query, which is completely irrelevant to how often a string actually appears on the web.
I looked at those links. They show random GitHub code samples which contain this model identifier. Even if K3 did train on those GitHub code samples, those are random code strings and K3 isn’t just reciting random pieces of code. When prefilled with "I am Claude", K3 answers with Anthropic's exact backend API identifiers, instead of the human conversational-style name "Opus 4.5".
If this was a result of scraping the open web and GitHub, other frontier models trained on GitHub data (like Qwen, Claude, GPT) would show a similar API model name when prompted. But they don't. Only K3 does this behavior, and it prefers to answer with the exact API tags like "claude-opus-4-5-20251101" and "claude-sonnet-4-20250514".
A LLM doesn't suddenly output a rare API model identifier when prompted with "I am Claude", just because it saw it in a .py file. No, it only does this because that identifier must have appeared rather frequently in the training data. Which is what happens when you distill Claude models, and don't properly clean your distillation data.
Plus, this isn't just spurious speculation that Moonshot is distilling Claude. Anthropic caught Moonshot as running "industrial-scale" distillation campaign from Claude: https://www.anthropic.com/news/detecting-and-preventing-dist...
To quote:
Moonshot (Kimi models) employed hundreds of fraudulent accounts spanning multiple access pathways. Varied account types made the campaign harder to detect as a coordinated operation. We attributed the campaign through request metadata, which matched the public profiles of senior Moonshot staff. In a later phase, Moonshot used a more targeted approach, attempting to extract and reconstruct Claude’s reasoning traces.
The operation targeted:
* Agentic reasoning and tool use
* Coding and data analysis
* Computer-use agent development
* Computer vision