X. Liang, “SEAP: Sparse Expert Activation Pruning Unlocks the Brainpower of Large Language Models”, AAAI, vol. 40, no. 38, pp. 31934–31942, Mar. 2026.