Clarify World Models Eval dataset attribution

README.md

@@ -64,9 +64,40 @@ The Hugging Face dataset viewer is configured through the dataset-card `configs`
 read `data/chunk-000/file-000.parquet` as the `default/train` split. The LeRobot videos
 and metadata files remain in the repository as sidecar assets for `LeRobotDataset`.
 
+## Creator And Maintainer
+
+World Models Eval is created and maintained by **Zaid Ghazal**.
+
+- Hugging Face: [zaid9876](https://huggingface.co/zaid9876)
+- GitHub: [ZaidGhazal/world-models-eval](https://github.com/ZaidGhazal/world-models-eval)
+- Dataset: [zaid9876/world-models-eval](https://huggingface.co/datasets/zaid9876/world-models-eval)
+
+This repository is not a mirror of the raw LIBERO release. It is a processed, normalized,
+LeRobotDataset-v3 dataset prepared for the World Models Eval study: fixed splits, held-out
+tasks, action/proprio normalization statistics, Hub viewer configuration, and sidecar videos
+organized for policy training, world-model training, simulator evaluation, and dream-rollout
+calibration.
+
 ## Source And Citation
 
-This dataset is converted from LIBERO demonstration HDF5 files. If you use it, cite LIBERO:
+The demonstrations come from the LIBERO benchmark, so LIBERO should be cited as the upstream
+source. If you use this processed dataset, please also cite or acknowledge **World Models Eval**
+and **Zaid Ghazal** for the dataset processing, normalization, splits, and release.
+
+Suggested citation for this processed dataset:
+
+```bibtex
+@misc{ghazal2026worldmodelseval,
+  title        = {World Models Eval: Processed LIBERO Demonstrations for World-Model-Based Robot Policy Evaluation},
+  author       = {Ghazal, Zaid},
+  year         = {2026},
+  howpublished = {Hugging Face dataset},
+  url          = {https://huggingface.co/datasets/zaid9876/world-models-eval},
+  note         = {LeRobotDataset v3 processed dataset with fixed splits and normalization statistics}
+}
+```
+
+Upstream LIBERO citation:
 
 ```bibtex
 @article{liu2023libero,