cognitivecomputations/SystemChat-2.0
Preview
β’
Updated
β’
87
β’
54
Note good to teach the model to follow system prompts. They are too specific though (and sometimes hard). We should add more generic system prompts in other samples of our dataset
Note filtered from https://huggingface.co./datasets/BAAI/Infinity-Instruct the 3M and 7M datasets boost HellaSwag, MMLU GSM8k..
Note Below H4 formats for the handbook trainings