configs:
- config_name: seed_42
data_files:
- split: train
path: data/seed_42/train.jsonl
- split: validation
path: data/seed_42/validation.jsonl
- split: test
path: data/seed_42/test.jsonl
- config_name: seed_1234
data_files:
- split: train
path: data/seed_1234/train.jsonl
- split: validation
path: data/seed_1234/validation.jsonl
- split: test
path: data/seed_1234/test.jsonl
- config_name: seed_2026
data_files:
- split: train
path: data/seed_2026/train.jsonl
- split: validation
path: data/seed_2026/validation.jsonl
- split: test
path: data/seed_2026/test.jsonl
CreativeEval
Paper-grouped multidimensional research-ideation evaluation data from CreativeEval.
Contents
The release contains 1,026 complete paper rows. Each row contains a human-written research-paper introduction, the raw reviewer score arrays for provenance, and four mean prediction targets: contribution_mean, soundness_mean, presentation_mean, and overall_score_mean.
All four targets are derived from the human reviewer scores released with the paper. The raw reviewer arrays are retained for auditability and should not be provided as predictor inputs.
Split organization
Each seed has paper-grouped train, validation, and test splits. One paper introduction remains in exactly one split. There is no OOD split because CreativeEval does not define a clean unseen venue or paper collection.
| Seed | Train rows | Validation rows | Test rows |
|---|---|---|---|
| 42 | 718 | 103 | 205 |
| 1234 | 718 | 103 | 205 |
| 2026 | 718 | 103 | 205 |
The paper-group counts are 718 train, 103 validation, and 205 test papers for every seed.
Provenance
Paper: https://aclanthology.org/2026.eacl-long.297/ Original code and data: https://github.com/lichun-19/creative_eval
Rebuild with:
python data/creativeeval_quality/build_huggingface_dataset.py \
--source-file /path/to/research_ideation.json