davidnichols-ops/

adaptive-operator-v4-dataset

Training data for adaptive-operator-v4, a Qwen3.5-9B model fine-tuned with custom control tokens for adaptive compute allocation in agentic workflows.

	Dataset Summary




	

Split Examples Format Size

SFT 4,992 OpenAI chat messages 8.8 MB

DPO 5,000 Chosen/rejected pairs 8.3 MB

Raw teacher responses 5,000 OpenAI chat messages 6.5 MB

Improved responses 4,992 OpenAI chat messages 13 MB

Preference pairs (raw) 5,000… See the full description on the dataset page: https://huggingface.co/datasets/davidnichols-ops/adaptive-operator-v4-dataset.