ROBORMBENCH
Explore score history ↗Evaluates paraphrase invariance in vision-language reward models for robotic learning using 2,390 real-robot trajectories, ground-truth progress labels, and 21,673 verified paraphrases spanning lexical, syntactic, and action-goal rewrites, measuring stability of predicted progress scores under semantically equivalent instructions.
- Released
- 2026-09-04
- Evaluates
- Robotics, Robotics & Embodied AI
- Openness
- unknown
- Importer
- Claire Radar
- Review status
- ai-name-audit-deferred
ROBORMBENCH paper
No reported scores are on record for this benchmark yet.