Unverified50% confidenceFactExact time
The paper's experiments show that current mainstream VLA policies perform poorly in zero-shot personalized scenarios: they can complete tasks, but it is difficult for them to follow preferences from instructions alone.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/16/2026
First Seen
Valid until: 1/14/2027
Sources
Related Entities
Related Claims
UnverifiedOne run fell below the most naive diff method due to a missing normalization step, demonstrating that preprocessing steps have a decisive impact on this type of task.68% similarUnverifiedREFACTOR-VLA directly confronts the core challenge of behavioral equivalence judgment that previous methods deliberately avoided.68% similarUnverifiedThere is currently no general answer for designing appropriate difficulty gradients and success thresholds for different types of tasks.68% similarUnverifiedIn VSArena's VLA track, the exact pose of the blocks is not passed to the policy model.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1041748API
curl https://kongchang.com/api/v1/knowledge/claims/1041748MCP
get_claim(id=1041748)