Nex-N2 Pro and Qwen Models Combined to Create Rio-3.5-Open
qwen
| Source: Mastodon | Original article
Rio's "homegrown" LLM is reportedly a merged model.
Rio de Janeiro's city government has been found to have misrepresented its "homegrown" large language model, Rio-3.5-Open-397B. As reported on GitHub, the model is approximately a 0.6 to 0.4 merge of Nex-N2_pro and Qwen, rather than an original creation. This revelation has significant implications for the transparency and credibility of AI development, particularly in the public sector.
The discovery matters because it highlights the potential for misrepresentation in AI research and development. If a government entity can pass off a merged model as its own, it raises questions about the integrity of AI research and the potential for similar instances of misrepresentation. This incident may also impact the trustworthiness of AI models developed by government agencies and the perceived value of open-source models like Nex-N2_pro.
As the story unfolds, it will be important to watch how Rio de Janeiro's city government responds to these allegations and whether they will take steps to rectify the situation. Additionally, the AI community will likely be monitoring the situation to see if similar instances of misrepresentation are uncovered, and how this incident will impact the development of AI models in the future.
Sources
Back to AIPULSEN