科研速览 · Science Skim继续刷下去 · Keep skimming →
◇ arXiv2026-08-21· cs.CL

Sparse Token Routing in Efficient Transformers

Sai Krishna Arthanari, JaeHyeong Chang, Chengzhe Sun, Siwei Lyu

原始摘要(英文原文)· Original abstract
Efficient-transformer research often motivates token pruning and adaptive computation with the claim that not all tokens require equal computational effort. We test this claim end to end using SEWN, a two-stream Transformer that routes tokens through either lightweight or full-capacity processing using a learned gate. Across our experiments, routing introduces negligible accuracy change relative to parameter-matched baselines, while the gate's token-importance signal depends critically on how it is learned. A static lexicon-seeded prior fails a counterfactual faithfulness test on BoolQ, whereas a fully contextual gate achieves highly significant separation ($p<10^{-10}$) on both evaluated tasks without changing task accuracy.
读原文 · Read the paper ↗

AI 追问PRO

登录后使用 AI 追问

讨论区

登录后参与讨论

相关论文 · Related

Sparse Token Routing in Efficient Transformers — 科研速览 Science Skim