Neeraj VarshneyAgneet ChatterjeeMihir ParmarChitta BaralInvestigating Acceleration of LLaMA Inference by Enabling Intermediate Layer Decoding via Instruction Tuning with 'LITE'.3656-36772024NAACL-HLT (Findings)https://doi.org/10.18653/v1/2024.findings-naacl.232conf/naacl/2024fdb/conf/naacl/naacl2024f.html#VarshneyCPB24streams/conf/naacl