tohio/

slm-synthetic-dpo-basic-arithmetic-qa

	SLM Synthetic DPO — basic_arithmetic_qa




	


	Summary

Synthetic preference-pair dataset containing chosen and rejected assistant responses for basic_arithmetic_qa.

Dataset type: preference optimization Total pairs: 6429 Signal: basic_arithmetic_qa Language: English

Category: direct_arithmetic Difficulty: 1 Template family: direct_addition Eval family: basic_arithmetic_qa

{ "id": "string"… See the full description on the dataset page: https://huggingface.co/datasets/tohio/slm-synthetic-dpo-basic-arithmetic-qa.