6 citations · 7 across the 11 of their papers we have counts for
Showing cs.CLShow all
2 papers · 1 filter
cs.CL2026
Target-Token Asymmetry Can Masquerade As Depth Specialization
Danlin Li, Miao Li, Xu Han +1
Layer-ablation studies often perturb early or late layers of a language model, compare the damage to prediction on two corpora, and read a larger early-layer effect on one corpus a…
cs.CL2026
Free Energy Mixer
Jiecheng Lu, Shihao Yang
Standard attention stores keys/values losslessly but reads them via a per-head convex average, blocking channel-wise selection. We propose the Free Energy Mixer (FEM): a free-energ…