allenai/Dolci-Think-DPO-32B
收藏Hugging Face2025-11-20 更新2025-12-20 收录
下载链接：
https://hf-mirror.com/datasets/allenai/Dolci-Think-DPO-32B
下载链接
链接失效反馈官方服务：
资源简介：
---
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
list:
- name: content
dtype: string
- name: role
dtype: string
- name: rejected
list:
- name: content
dtype: string
- name: role
dtype: string
- name: chosen_model
dtype: string
- name: rejected_model
dtype: string
- name: dataset
dtype: string
- name: prompt_id
dtype: string
- name: preference_type
dtype: string
splits:
- name: train
num_bytes: 4488651554
num_examples: 200000
download_size: 1882695224
dataset_size: 4488651554
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: odc-by
---
# Dolci Think DPO Mixture
This dataset is licensed under ODC-BY. It is intended for research and educational use in accordance with Ai2's [Responsible Use Guidelines](https://allenai.org/responsible-use).
The Dolci Think DPO mixture was used to preference tune Olmo 3 Think 32B. It contains 200,000 preference pairs created with the preference heuristic described in [Delta Learning](https://arxiv.org/abs/2507.06187) (Geng et al. 2025).
提供机构：
allenai
