ERQA

From the catalogue. Listed by one researcher from one source page.

• id: erqa
• kind: language and reasoning annotation
• robot or device: real-world robotics-focused images (class: sensor only)
• how collected: curated multiple-choice questions
• size as stated: 400 examples
• hours: not stated | hours only claimed: none | episodes: not stated
• year: 2025
• organisation: Google DeepMind
• licence: CC-BY-4.0 (class: open)
• access: open (class: open)
• link (dataset): https://github.com/embodiedreasoning/ERQA
• note: Released with the Gemini Robotics report; TFRecord format.

Confirm the licence at the link before relying on it.