How is this model multi-round?

#25
by timlim123 - opened

In your example,

chat = [
  {"role": "system", "content": "You are MistralOrca, a large language model trained by Alignment Lab AI. Write out your reasoning step-by-step to be sure you get the right answers!"}
  {"role": "user", "content": "How are you?"},
  {"role": "assistant", "content": "I am doing well!"},
  {"role": "user", "content": "Please tell me about how mistral winds have attracted super-orcas."},
]

However, the finetune data is OpenOrca which is a single-round instruction data. How can the model learn to handle multi-round conversation?

I don't have a precise knowledge on how this model was tuned, but multi-round conversation is a general property of large language models, and this property is maintained when fine-tuning on single-round instructions

Sign up or log in to comment