phi-monsterapache-2.0

Deconfounded negation set (RoboCasa)

A dataset of 517 episodes from RoboCasa featuring negation-grounded language instructions where positive and negated instructions (e.g., 'the apple' vs 'not the apple') select different objects, designed to deconfound negation from keyword matching.

Downloads3
Episodes517

Why This Matters for Physical AI

This dataset addresses the critical problem of language grounding in robotic manipulation by providing controlled negation examples that help models learn robust semantic understanding beyond simple keyword matching.

Technical Profile

Modalities
rgblanguage
Environment
simulation
Task Types
manipulationgrounding
Episodes
517
Data Format
LeRobot
Annotation Types
language_instructions
License
apache-2.0
Part of the RoboCasa family

Access

Need custom rgb data?

Claru builds purpose-built datasets for simulation applications with dense human annotations and quality assurance.

Request a Sample Pack

Related Datasets