Tracking instruction-tuned LLM openness. Paper: Liesenfeld, Andreas, Alianda Lopez, and Mark Dingemanse. 2023. “Opening up ChatGPT: Tracking Openness, Transparency, and Accountability in Instruction-Tuned Text Generators.” In Proceedings of the 5th International Conference on Conversational User Interfaces. doi:10.1145/3571884.3604316.
8B and 16B chat models found through Alpaca Eval
https://huggingface.co/Nanbeige/Nanbeige2-8B-Chat
some code available on github, no model card, no data sheet, no info on training and SFT data