一作ACL 2026Findings· Findings of ACL
P ib E - MPP : A Play-it-by-Ear Masking Performance Plug-in for LLM s
Mengwei Wang · Simin Niu · Xun Liang · Yuefeng Ma · Sensen Zhang · Jiawei Yang · Shichao Song · Hanyu Wang · Huayi Lai
摘要
Treating random masking as a performance plug-in for large language models (LLMs) offers three advantages: low coupling to the task, the model, and training resources. However, the critical drawback is that its gains are highly stochastic. Motivated by this, we propose play-it-by-ear masking perf...