policy
Datasets
All datasets matching “policy”policy-docs
Public Policy at Hugging Face
AI Policy at Hugging Face is a multidisciplinary and cross-organizational workstream. Instead of being part of a vertical communications or global affairs organization, our policy work is rooted in the expertise of our many researchers and developers, from Ethics and Society Regulars and legal team to machine learning engineers working on healthcare, art, and evaluations.
What we work on is informed by our Hugging Face community needs and experiences… See the full description on the dataset page: https://huggingface.co/datasets/huggingface/policy-docs.populace-us
populace-us
The populace-built US population: a calibrated synthetic microdataset for
PolicyEngine-US, built by the
populace stack entirely from
primary sources — the enhanced CPS appears only as the benchmark this file is
scored against, never as a build input. It loads anywhere the enhanced CPS
loads (an API-compatible alternative population), with its own calibrated
weights — and its own strengths and gaps, both documented below.
Load it
pip install… See the full description on the dataset page: https://huggingface.co/datasets/policyengine/populace-us.LIBERO-Cosmos-Policy
LIBERO-Cosmos-Policy
Dataset Description
LIBERO-Cosmos-Policy is a modified version of the LIBERO simulation benchmark dataset, created as part of the Cosmos Policy project. This is the dataset used to train the Cosmos-Policy-LIBERO-Predict2-2B checkpoint.
Key Modifications
Our modifications include the following:
Higher-resolution images: Images are saved at 256×256 pixels (vs. 128×128 in the original).
No-op actions filtering: Transitions with "no-op" (zero)… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/LIBERO-Cosmos-Policy.RoboCasa-Cosmos-Policy
RoboCasa-Cosmos-Policy
Dataset Description
RoboCasa-Cosmos-Policy is a modified version of the RoboCasa simulation benchmark dataset, created as part of the Cosmos Policy project. This is the dataset used to train the Cosmos-Policy-RoboCasa-Predict2-2B checkpoint.
Key Modifications
Our modifications include the following:
Higher-resolution images: Images are saved at 224×224 pixels (vs. 128×128 in the original).
No-op actions filtering: Transitions with… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/RoboCasa-Cosmos-Policy.diffusion_policy_robocasa_activations_latest_chkpt
Diffusion Policy — RoboCasa Activations (latest checkpoint)
Per-step, per-episode activation traces collected from a DiffusionTransformerHybridImagePolicy (the diffusion_policy library) rolled out on RoboCasa benchmark tasks. Captured with collect_activations_robocasa.py at the latest training checkpoint.
These traces are the input expected by the conceptor / SAE steering pipelines under diffusion_policy/experiments/robocasa_steering/ and diffusion_policy/experiments/sae/ — see… See the full description on the dataset page: https://huggingface.co/datasets/brandonyang/diffusion_policy_robocasa_activations_latest_chkpt.music-off-policy-evaluation-benchmark
Music Off-Policy Evaluation Dataset
Music Off-Policy Evaluation Dataset is a dataset designed for Off-Policy Evaluation (OPE) research. It contains logged interactions from the home page of Amazon Music.
Use cases:
Benchmarking OPE estimators
Evaluating counterfactual ranking policies offline
License
Music Off-Policy Evaluation Benchmark © 2026 by Amazon is licensed under Creative Commons Attribution-NonCommercial 4.0 International.… See the full description on the dataset page: https://huggingface.co/datasets/amazon/music-off-policy-evaluation-benchmark.