A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulation

Recommended to me by Krish Mehta.

Project page:

• https://toyotaresearchinstitute.github.io/lbm1/

They look at different scaling laws for robotics.

• Uses a Diffusion Transformer

Takeaways:

• Robotics is extremely noisy, in this paper, they use statistical rigor to get conclusive results
• Some talks on normalization

Citation

(Team et al., 2025)

Team, T. L., Barreiros, J., Beaulieu, A., Bhat, A., Cory, R., Cousineau, E., Dai, H., Fang, C.-H., Hashimoto, K., Irshad, M. Z., Itkina, M., Kuppuswamy, N., Lee, K.-H., Liu, K., McConachie, D., McMahon, I., Nishimura, H., Phillips-Grafflin, C., Richter, C., … Tedrake, R. (2025). A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulation. https://arxiv.org/pdf/2507.05331