A compact evaluation dataset and harness for testing agentic AI on safety-critical robotics tasks. Includes multimodal episodes in parquet format, task-specific eval scripts (gauge reading, human safety monitoring, VLA estimators), and TFDS/Hugging Face integration for reproducible safety evaluations.
[next-mdx-remote-client] error compiling MDX: Unexpected character `1` (U+0031) before name, expected a character that can start a name, such as a letter, `$`, or `_` More information: https://mdxjs.com/docs/troubleshooting-mdx
Provides source-grounded, long-form instruction data for defensive cybersecurity assistants across 17 domains. Mixed-source licensing and reusable reference appendices require care when training on or redistributing the data.
Provides fine-grained language labels across complete bimanual robot trajectories, enabling step-by-step policy learning from 40,543 demonstrations and 533,913 human-written subtasks.
Pairs natural-language scene prompts with complete Three.js HTML/JavaScript responses, generated reasoning traces, and optional rendered previews. The examples span varied subjects and complexity levels for training or evaluating 3D code-generation models.
