Supervised fine-tuning corpus for the Phytomni multi-agent system. The corpus is used to fine-tune both Phytomni LLMs — Phyto-Chatbot and Phyto-Reasoner — after continual pre-training on Phytomni-Pretrain.
The release contains 20,982 examples, split evenly between the two models: 10,491 for Phyto-Reasoner and 10,491 for Phyto-Chatbot. Each per-model half mixes plant-science domain examples with general-domain examples (math, programming… See the full description on the dataset page:
https://huggingface.co/datasets/Phytomni/Phytomni-SFT.
