Firehose

16 SEP 2026 · Paper

ActionPiece: Rethinking Action Tokenization for Autoregressive Vision-Language-Action Models

This paper improves autoregressive vision-language-action models by creating a new method for action tokenization that better preserves the relationships between actions, allowing the model to perform more accurately in different contexts. Practitioners might care about this because it could lead to more reliable and generalizable vision-language-action models.

vision-language-action models action tokenization physical rank consistency quantization representation learning

Browse by tag

ActionPiece: Rethinking Action Tokenization for Autoregressive Vision-Language-Action Models