REFLECT / RoboFail

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

• id: reflect-robofail
• kind: language and reasoning annotation
• robot or device: simulated and real robot failure executions (class: single arm)
• how collected: robot executions with injected failures
• size as stated: 10 different tasks (per AweRoFaM list)
• hours: not stated | hours only claimed: none | episodes: not stated
• year: 2023
• organisation: Columbia University
• licence: not stated (class: not stated)
• access: open (class: open)
• link (dataset): https://robot-reflect.github.io/
• other names: RoboFail (REFLECT)
• note: Multi-sensory failure explanation dataset.

Confirm the licence at the link before relying on it.