# FFN-SkipLLM: A Hidden Gem for Autoregressive Decoding with Adaptive Feed Forward Skipping

arXiv: https://arxiv.org/abs/2404.03865 · HTML: https://mlsanity.com/paper/2404.03865/

## Key facts

- Cited by 23 papers in the map (data through 14 Oct 2026).
- Placed in Speculative decoding, one of 12 directions in Efficient LLM inference and serving.
- Builds on 55 of its 65 references that are in the map.
- First posted on arXiv 5 Apr 2024.

## Placed in

- Efficient LLM inference and serving / [Speculative decoding](https://mlsanity.com/inference/speculative-decoding/)

## Builds on (most cited first, 30 of 55)

- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Neural Machine Translation by Jointly Learning to Align and Translate](https://mlsanity.com/paper/1409.0473/) (2014)
- [SQuAD: 100,000+ Questions for Machine Comprehension of Text](https://mlsanity.com/paper/1606.05250/) (2016)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension](https://mlsanity.com/paper/1705.03551/) (2017)
- [The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks](https://mlsanity.com/paper/1803.03635/) (2018)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers](https://mlsanity.com/paper/2210.17323/) (2022)
- [SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot](https://mlsanity.com/paper/2301.00774/) (2023)
- [Abstractive Text Summarization Using Sequence-to-Sequence RNNs and Beyond](https://mlsanity.com/paper/1602.06023/) (2016)
- [SNIP: Single-shot Network Pruning based on Connection Sensitivity](https://mlsanity.com/paper/1810.02340/) (2018)
- [LISA: Reasoning Segmentation via Large Language Model](https://mlsanity.com/paper/2308.00692/) (2023)
- [In-Context Retrieval-Augmented Language Models](https://mlsanity.com/paper/2302.00083/) (2023)
- [FlexGen: High-Throughput Generative Inference of Large Language Models with a Single GPU](https://mlsanity.com/paper/2303.06865/) (2023)
- [VisionLLM: Large Language Model is also an Open-Ended Decoder for Vision-Centric Tasks](https://mlsanity.com/paper/2305.11175/) (2023)
- [Is ChatGPT a General-Purpose Natural Language Processing Task Solver?](https://mlsanity.com/paper/2302.06476/) (2023)
- [Multi-News: a Large-Scale Multi-Document Summarization Dataset and Abstractive Hierarchical Model](https://mlsanity.com/paper/1906.01749/) (2019)
- [ShortGPT: Layers in Large Language Models are More Redundant Than You Expect](https://mlsanity.com/paper/2403.03853/) (2024)
- [SpQR: A Sparse-Quantized Representation for Near-Lossless LLM Weight Compression](https://mlsanity.com/paper/2306.03078/) (2023)
- [Chameleon: Plug-and-Play Compositional Reasoning with Large Language Models](https://mlsanity.com/paper/2304.09842/) (2023)
- [DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference](https://mlsanity.com/paper/2004.12993/) (2020)
- [LLM-QAT: Data-Free Quantization Aware Training for Large Language Models](https://mlsanity.com/paper/2305.17888/) (2023)
- [The Lottery Ticket Hypothesis for Pre-trained BERT Networks](https://mlsanity.com/paper/2007.12223/) (2020)
- [Exploring the Potential of Large Language Models (LLMs) in Learning on Graphs](https://mlsanity.com/paper/2307.03393/) (2023)
- [DynaBERT: Dynamic BERT with Adaptive Width and Depth](https://mlsanity.com/paper/2004.04037/) (2020)
- [LLM-grounded Diffusion: Enhancing Prompt Understanding of Text-to-Image Diffusion Models with Large Language Models](https://mlsanity.com/paper/2305.13655/) (2023)
- [Depth-Adaptive Transformer](https://mlsanity.com/paper/1910.10073/) (2019)
- [FastBERT: a Self-distilling BERT with Adaptive Inference Time](https://mlsanity.com/paper/2004.02178/) (2020)

## Built on by (most cited first, 23 of 23)

- [PointVLA: Injecting the 3D World into Vision-Language-Action Models](https://mlsanity.com/paper/2503.07511/) (2025)
- [SWIFT: On-the-Fly Self-Speculative Decoding for LLM Inference Acceleration](https://mlsanity.com/paper/2410.06916/) (2024)
- [Tracing Representation Progression: Analyzing and Enhancing Layer-Wise Similarity](https://mlsanity.com/paper/2406.14479/) (2024)
- [A deeper look at depth pruning of LLMs](https://mlsanity.com/paper/2407.16286/) (2024)
- [SwiftKV: Fast Prefill-Optimized Inference with Knowledge-Preserving Model Transformation](https://mlsanity.com/paper/2410.03960/) (2024)
- [Routing Experts: Learning to Route Dynamic Experts in Multi-modal Large Language Models](https://mlsanity.com/paper/2407.14093/) (2024)
- [AdaSkip: Adaptive Sublayer Skipping for Accelerating Long-Context LLM Inference](https://mlsanity.com/paper/2501.02336/) (2025)
- [QuantMoE-Bench: Examining Post-Training Quantization for Mixture-of-Experts](https://mlsanity.com/paper/2406.08155/) (2024)
- [NoVo: Norm Voting off Hallucinations with Attention Heads in Large Language Models](https://mlsanity.com/paper/2410.08970/) (2024)
- [FBS: Modeling Native Parallel Reading inside a Transformer](https://mlsanity.com/paper/2601.21708/) (2026)
- [BlockLLM: Multi-tenant Finer-grained Serving for Large Language Models](https://mlsanity.com/paper/2404.18322/) (2024)
- [MemoryLLM: Plug-n-Play Interpretable Feed-Forward Memory for Transformers](https://mlsanity.com/paper/2602.00398/) (2026)
- [Advancing Decoding Strategies: Enhancements in Locally Typical Sampling for LLMs](https://mlsanity.com/paper/2506.05387/) (2025)
- [RAP: Runtime Adaptive Pruning for LLM Inference](https://mlsanity.com/paper/2505.17138/) (2025)
- [DASH: Input-Aware Dynamic Layer Skipping for Efficient LLM Inference with Markov Decision Policies](https://mlsanity.com/paper/2505.17420/) (2025)
- [Prada: Black-Box LLM Adaptation with Private Data on Resource-Constrained Devices](https://mlsanity.com/paper/2503.14932/) (2025)
- [High-Layer Attention Pruning with Rescaling](https://mlsanity.com/paper/2507.01900/) (2025)
- [Confidence-Modulated Speculative Decoding for Large Language Models](https://mlsanity.com/paper/2508.15371/) (2025)
- [Fast Forward: Accelerating LLM Prefill with Predictive FFN Sparsity](https://mlsanity.com/paper/2602.00397/) (2026)
- [FiRST: Finetuning Router-Selective Transformers for Input-Adaptive Latency Reduction](https://mlsanity.com/paper/2410.12513/) (2024)
- [QTALE: Quantization-Robust Token-Adaptive Layer Execution for LLMs](https://mlsanity.com/paper/2602.10431/) (2026)
- [TIDE: Every Layer Knows the Token Beneath the Context](https://mlsanity.com/paper/2605.06216/) (2026)
- [BALANCE: Hybrid Autoregressive-Speculative LLM Inference at the Network Edge](https://mlsanity.com/paper/2608.05926/) (2026)

## Abstract (arXiv)

Autoregressive Large Language Models (e.g., LLaMa, GPTs) are omnipresent achieving remarkable success in language understanding and generation. However, such impressive capability typically comes with a substantial model size, which presents significant challenges for autoregressive token-by-token generation. To mitigate computation overload incurred during generation, several early-exit and layer-dropping strategies have been proposed. Despite some promising success due to the redundancy across LLMs layers on metrics like Rough-L/BLUE, our careful knowledge-intensive evaluation unveils issues such as generation collapse, hallucination of wrong facts, and noticeable performance drop even at the trivial exit ratio of 10-15% of layers. We attribute these errors primarily to ineffective handling of the KV cache through state copying during early-exit. In this work, we observed the saturation of computationally expensive feed-forward blocks of LLM layers and proposed FFN-SkipLLM, which is a novel fine-grained skip strategy of autoregressive LLMs. More specifically, FFN-SkipLLM is an input-adaptive feed-forward skipping strategy that can skip 25-30% of FFN blocks of LLMs with marginal change in performance on knowledge-intensive generation tasks without any requirement to handle KV cache. Our extensive experiments and ablation across benchmarks like MT-Bench, Factoid-QA, and variable-length text summarization illustrate how our simple and ease-at-use method can facilitate faster autoregressive decoding.