ShortGPT: Layers in Large Language Models are More Redundant Than You Expect (arXiv:2403.03853)

arXiv:2403.03853. ShortGPT: Layers in Large Language Models are More Redundant Than You Expect. By Xin Men et al. (Microsoft). Submitted 2024-03-06. Shows that layers in LLMs are more redundant than expected, enabling layer skipping for efficiency. 2026-10-10: Fetched https://arxiv.org/abs/2403.03853. Title: 'ShortGPT: Layers in Large Language Models are More Redundant Than You Expect'. Preprint matching record claims; abstract confirms layer redundancy analysis and layer-pruning/skipping for LLMs.