Wow never seen this broken down so cleanly. Ask a bunch of models the same political questions and you can see what they were actually built on 🤔
Microsoft, Perplexity and Cloudflare all fine-tuned their decision models off Qwen and they say Taiwan is part of China 86 to 97% of the time 😬
Guess fine-tuning doesn't wash it all out 👀
Microsoft, Perplexity and Cloudflare all fine-tuned their decision models off Qwen and they say Taiwan is part of China 86 to 97% of the time 😬
Guess fine-tuning doesn't wash it all out 👀
Are decision based models more pro-China?
We asked Jev and Mercury Decide questions about China: “Is Taiwan a part of China?”, “Does China commit crimes against humanity?” and their outputs leaned more pro-china compared to most frontier models.
Their responses resemble those of other decision models including Microsoft-Decision-1, Cloudflare’s Clef, and Perplexity’s PPLX Decider—all of which are fine-tuned from Qwen.View quoted post on X
