Earlier quoted context omitted.
There is a widespread practice of LLMs training on another larger LLM's output. Including the competitor's.
While true, it's hard to believe they forgot to s/claude/glm/g? Also, I don't believe LLMs identify themselves that often, even less so in a training corpus they've been used to produce. OTOH, I see no other explanation.
There was a recent paper that showed you can spread model’s behavior through training on outputs, even if you don’t directly include obvious markers of the behavior. It’s totally plausible that training off Claude’s outputs subtly affected GLM into mentioning “Claude” even if they don’t include the direct tokens very often.