A politically biased model released by SCAM-AI
To be honest, I haven't heard of your company before (since your company appears to be a scam company).
However, I have some serious concerns about your approach. It appears you're leveraging a Chinese open‐source model and training it exclusively on politically driven data. This strategy raises questions about the objectivity and reliability of your LLM.
In light of this, I think "Political-AI" would be a more fitting name for your company.
After further fine-tuning R1, the released model's performance on several public benchmarks is even worse than that of the original R1. Moreover, I haven’t come across any notable publications from your organization. I remain puzzled as to why your company continues to operate. Do you really work on AI, or is your company just a scam?
