README.md

---
license: other
license_name: multiple-licenses
license_link: https://huggingface.co/ActiveUltraFeedback/combined/blob/main/LICENSES.md
task_categories:
- text-generation
language:
- en
tags:
- preference-data
- active-learning
- RLHF
- DPO
- reward-modeling
arxiv: 2603.09692
pretty_name: ActiveUltraFeedback — Combined
configs:
- config_name: full
  data_files:
  - split: train
    path: full/train-*.parquet
- config_name: rm_maxminlcb
  data_files:
  - split: train
    path: rm_maxminlcb/train-*.parquet
- config_name: rm_dts
  data_files:
  - split: train
    path: rm_dts/train-*.parquet
- config_name: rm_infomax
  data_files:
  - split: train
    path: rm_infomax/train-*.parquet
- config_name: rm_deltaucb
  data_files:
  - split: train
    path: rm_deltaucb/train-*.parquet
- config_name: rm_drts
  data_files:
  - split: train
    path: rm_drts/train-*.parquet
- config_name: dpo_maxminlcb
  data_files:
  - split: train
    path: dpo_maxminlcb/train-*.parquet
- config_name: dpo_dts
  data_files:
  - split: train
    path: dpo_dts/train-*.parquet
- config_name: dpo_infomax
  data_files:
  - split: train
    path: dpo_infomax/train-*.parquet
- config_name: dpo_deltaucb
  data_files:
  - split: train
    path: dpo_deltaucb/train-*.parquet
- config_name: dpo_drts
  data_files:
  - split: train
    path: dpo_drts/train-*.parquet
---

# ActiveUltraFeedback — Combined

This is a preference dataset of 140k samples generated for the paper **ActiveUltraFeedback: Efficient Preference Data Generation using Active Learning** ([Melikidze et al., 2026](https://arxiv.org/abs/2603.09692)). 

The prompts are from a combination of the version of UltraFeedback (Cui et al., 2023) that was released by AllenAI (at [allenai/ultrafeedback_binarized_cleaned](https://huggingface.co/datasets/allenai/ultrafeedback_binarized_cleaned)) and [Skywork Reward Preference 80k v0.2](https://huggingface.co/datasets/Skywork/Skywork-Reward-Preference-80K-v0.2) (Liu et al., 2024). The response pairs were generated with the ActiveUltraFeedback pipeline, which calls a large pool of open-weight LLMs to first generate candidate responses, then uses various active selection strategies to choose which response pair per prompt to select for preference annotation by an LLM-as-a-Judge. 

## Configs

Each of the 11 configs is a `train` split. The largest is called `full`, and it contains all 30 responses that we generated for each prompt. The remaining 10 splits have only two responses per prompt, and the config names are of this form: `<preference_optimization_method>_<response_pair_selection_method>`.

**Preference optimization method** (2 options):
- `rm` — Reward Modeling
- `dpo` — Direct Preference Optimisation

**Response pair selection method** (5 options): The first three are based on existing works, and the final two are novel. 
- `infomax` — InfoMax ([Saha, 2021](https://papers.nips.cc/paper_files/paper/2021/hash/fc3cf452d3da8402bebb765225ce8c0e-Abstract.html)) prioritizes pure exploration by selecting the response pair with the highest joint uncertainty, regardless of the predicted reward quality.
- `dts` — DTS (Double Thompson Sampling) ([Wu & Lin, 2016](https://papers.nips.cc/paper_files/paper/2016/hash/9de6d14fff9806d4bcd1ef555be766cd-Abstract.html)) addresses the exploration-exploitation trade-off by drawing two independent samples from the reward posterior and selecting the responses that maximize them.
- `maxminlcb` — MaxMinLCB ([Pásztor et al., 2024](https://proceedings.neurips.cc/paper_files/paper/2024/hash/1646e34971facbcda3727d1dc28ab635-Abstract-Conference.html)) considers pairwise Lower Confidence Bounds (LCBs) and selects the response pair where the first response maximizes the minimum LCB against all responses, and the second response minimizes the LCB against the first response.
- `deltaucb` — DeltaUCB identifies pairs with the largest optimistic quality difference, i.e., the pair that maximizes the probability that the first response is preferred over the second response in the best-case scenario.
- `drts` — DRTS (Double Reversed Thompson Sampling) selects one response that maximizes and another that minimizes their respective samples from the reward posterior. This strategy explicitly targets pairs with a significant delta in quality, while the underlying stochastic sampling preserves exploration and diversity.

## Features

The features for the `full` config are listed below:

| Field | Type | Description |
|-------|------|-------------|
| `prompt_id` | string | Unique prompt identifier |
| `prompt` | string | Input prompt |
| `prompt_licenses` | list of strings | Licenses applicable to the input prompt |
| `source` | string | Where the original dataset obtained the prompt from. |
| `completions` | list of dict | Each dict contains a model name, the response generated by the model (along with the system prompt and messages for generation), and how good the response is on a given principle according to an LLM Judge. | 

The remaining 10 configs have these features:

| Field | Type | Description |
|-------|------|-------------|
| `prompt_id` | string | Unique prompt identifier |
| `prompt` | string | Input prompt |
| `prompt_licenses` | list of strings | Licenses applicable to the input prompt |
| `chosen` | list of `{role, content}` | Preferred response |
| `chosen_model` | string | Model that generated the chosen response |
| `chosen_score` | float | Reward model score for the chosen response |
| `chosen_licenses` | list of strings | Licenses applicable to the chosen response |
| `rejected` | list of `{role, content}` | Dispreferred response |
| `rejected_model` | string | Model that generated the rejected response |
| `rejected_score` | float | Reward model score for the rejected response |
| `rejected_licenses` | list of strings | Licenses applicable to the rejected response |

## Licenses

The prompts and responses are derived from third-party sources under their respective licenses. We have annotated every prompt and response with the licenses that may apply. Review all constituent licenses before use. See [LICENSES.md](LICENSES.md) for the full list.

## Citation

If you use this dataset, please cite:

```bibtex
@misc{melikidze2026activeultrafeedbackefficientpreferencedata,
      title={ActiveUltraFeedback: Efficient Preference Data Generation using Active Learning}, 
      author={Davit Melikidze and Marian Schneider and Jessica Lam and Martin Wertich and Ido Hakimi and Barna Pásztor and Andreas Krause},
      year={2026},
      eprint={2603.09692},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2603.09692}, 
}
```