README.md

metadata

dataset_info:
  features:
    - name: prompt
      dtype: string
    - name: chosen
      dtype: string
    - name: rejected
      dtype: string
    - name: lenTokens_chosen
      dtype: int64
    - name: lenTokens_rejected
      dtype: int64
  splits:
    - name: train
      num_bytes: 362783229
      num_examples: 120257
  download_size: 145437391
  dataset_size: 362783229
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: apache-2.0
task_categories:
  - text-generation
language:
  - es
  - en
size_categories:
  - 100K<n<1M

https://huggingface.co/datasets/NickyNicky/function_call_orpo_sft_phi3_chatML_only
https://huggingface.co/datasets/NickyNicky/oasst2_orpo_mix_tokenizer_phi_3_v1