# Softpick: No Attention Sink, No Massive Activations with Rectified Softmax

arXiv: https://arxiv.org/abs/2504.20966 · HTML: https://mlsanity.com/paper/2504.20966/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- Builds on 35 of its 59 references that are in the map.
- First posted on arXiv 29 Apr 2025.

## Builds on (most cited first, 30 of 35)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [Neural Machine Translation by Jointly Learning to Align and Translate](https://mlsanity.com/paper/1409.0473/) (2014)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [PIQA: Reasoning about Physical Commonsense in Natural Language](https://mlsanity.com/paper/1911.11641/) (2019)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [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)
- [GLU Variants Improve Transformer](https://mlsanity.com/paper/2002.05202/) (2020)
- [SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models](https://mlsanity.com/paper/2211.10438/) (2022)
- [The LAMBADA dataset: Word prediction requiring a broad discourse context](https://mlsanity.com/paper/1606.06031/) (2016)
- [LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scale](https://mlsanity.com/paper/2208.07339/) (2022)
- [Quantifying Attention Flow in Transformers](https://mlsanity.com/paper/2005.00928/) (2020)
- [Crowdsourcing Multiple Choice Science Questions](https://mlsanity.com/paper/1707.06209/) (2017)
- [SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning](https://mlsanity.com/paper/2012.09852/) (2020)
- [When Attention Sink Emerges in Language Models: An Empirical View](https://mlsanity.com/paper/2410.10781/) (2024)
- [Quantizable Transformers: Removing Outliers by Helping Attention Heads Do Nothing](https://mlsanity.com/paper/2306.12929/) (2023)
- [Why do LLMs attend to the first token?](https://mlsanity.com/paper/2504.02732/) (2025)
- [TorchTitan: One-stop PyTorch native solution for production ready LLM pre-training](https://mlsanity.com/paper/2410.06511/) (2024)
- [FP8-LM: Training FP8 Large Language Models](https://mlsanity.com/paper/2310.18313/) (2023)
- [Scaling FP8 training to trillion-token LLMs](https://mlsanity.com/paper/2409.12517/) (2024)
- [Active-Dormant Attention Heads: Mechanistically Demystifying Extreme-Token Phenomena in LLMs](https://mlsanity.com/paper/2410.13835/) (2024)
- [Training-Free Activation Sparsity in Large Language Models](https://mlsanity.com/paper/2408.14690/) (2024)
- [Scalable-Softmax Is Superior for Attention](https://mlsanity.com/paper/2501.19399/) (2025)
- [Agile-Quant: Activation-Guided Quantization for Faster Inference of LLMs on the Edge](https://mlsanity.com/paper/2312.05693/) (2023)
- [BitNet v2: Native 4-bit Activations with Hadamard Transformation for 1-bit LLMs](https://mlsanity.com/paper/2504.18415/) (2025)
- [LongHeads: Multi-Head Attention is Secretly a Long Context Processor](https://mlsanity.com/paper/2402.10685/) (2024)

## Built on by (most cited first, 22 of 22)

- [Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink-Free](https://mlsanity.com/paper/2505.06708/) (2025)
- [RADLADS: Rapid Attention Distillation to Linear Attention Decoders at Scale](https://mlsanity.com/paper/2505.03005/) (2025)
- [Unveiling Super Experts in Mixture-of-Experts Large Language Models](https://mlsanity.com/paper/2507.23279/) (2025)
- [Mitigating Attention Sinks and Massive Activations in Audio-Visual Speech Recognition with LLMs](https://mlsanity.com/paper/2510.22603/) (2025)
- [Attention Sinks Are Provably Necessary in Softmax Transformers: Evidence from Trigger-Conditional Tasks](https://mlsanity.com/paper/2603.11487/) (2026)
- [Attention Sink in Transformers: A Survey on Utilization, Interpretation, and Mitigation](https://mlsanity.com/paper/2604.10098/) (2026)
- [Hidden Dynamics of Massive Activations in Transformer Training](https://mlsanity.com/paper/2508.03616/) (2025)
- [On the Nature of Attention Sink that Shapes Decoding Strategy in Omni-LLMs](https://mlsanity.com/paper/2603.14337/) (2026)
- [Threshold Differential Attention for Sink-Free, Ultra-Sparse, and Non-Dispersive Language Modeling](https://mlsanity.com/paper/2601.12145/) (2026)
- [Attention Sink Forges Native MoE in Attention Layers: Sink-Aware Training to Address Head Collapse](https://mlsanity.com/paper/2602.01203/) (2026)
- [OScaR: The Occam's Razor for Extreme KV Cache Quantization in LLMs and Beyond](https://mlsanity.com/paper/2605.19660/) (2026)
- [Surgery: Mitigating Harmful Fine-Tuning for Large Language Models via Attention Sink](https://mlsanity.com/paper/2602.05228/) (2026)
- [EarlyTom: Early Token Compression Completes Fast Video Understanding](https://mlsanity.com/paper/2605.30010/) (2026)
- [Attention Sinks Induce Gradient Sinks: Massive Activations as Gradient Regulators in Transformers](https://mlsanity.com/paper/2603.17771/) (2026)
- [A Single Layer to Explain Them All:Understanding Massive Activations in Large Language Models](https://mlsanity.com/paper/2605.08504/) (2026)
- [Attention Sinks and Outliers in Attention Residuals](https://mlsanity.com/paper/2605.17887/) (2026)
- [FLUID: Continuous-Time Hyperconnected Sparse Transformer for Sink-Free Learning](https://mlsanity.com/paper/2605.04421/) (2026)
- [Disentangling Semantic Attention from Structural Bias in the Attention Manifold](https://mlsanity.com/paper/2607.24017/) (2026)
- [From Attention Masks to Inert Zero-Vector Tokens: OAttention and O-Closure for Token Dynamics](https://mlsanity.com/paper/2608.21174/) (2026)
- [Do New Attention Mechanisms Actually Fix Attention Sinks at Million-Token Context?](https://mlsanity.com/paper/2609.08574/) (2026)
- [Abstention and Noise Filtering: Two Missing Primitives of Softmax Attention](https://mlsanity.com/paper/2609.22005/) (2026)
- [Underscoring the Problem: Why Softpick Fails at Initialization](https://mlsanity.com/paper/2610.05488/) (2026)

## Abstract (arXiv)

We introduce softpick, a rectified, not sum-to-one, drop-in replacement for softmax in transformer attention mechanisms that eliminates attention sink and massive activations. Our experiments with 340M and 1.8B parameter models demonstrate that softpick achieves 0% sink rate consistently. The softpick transformers produce hidden states with significantly lower kurtosis and creates sparse attention maps. Quantized models using softpick outperform softmax on standard benchmarks, with a particularly pronounced advantage at lower bit precisions. Our analysis and discussion shows how softpick has the potential to open new possibilities for quantization, low-precision training, sparsity optimization, pruning, and interpretability. Our code: https://github.com/zaydzuhri/softpick-attention