Dataset for fine-tuning LoRA adapters to improve agent tool selection. Extracted from real Hermes Agent sessions.
Each line is a JSON object:
{
"messages": [
{"role": "system", "content": "..."}… See the full description on the dataset page:
https://huggingface.co/datasets/DrVon/hermes-tool-selector.