Smarter, Not Harder: Training-Free Adaptive Computation for Transformers.

Romain Stora, Jaeseong Lee, Seung-won Hwang

Browse the full ACL paper archive.