In-Context Linear Regression Demystified: Training Dynamics and Mechanistic Interpretability of Multi-Head Softmax Attention
Fuente:
arXiv
Saved in:
| Main Authors: | He, Jianliang, Pan, Xintian, Chen, Siyu, Yang, Zhuoran |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Similar Items
Training Dynamics of Multi-Head Softmax Attention for In-Context Learning: Emergence, Convergence, and Optimality
by: Chen, Siyu, et al.
Published: (2024)
by: Chen, Siyu, et al.
Published: (2024)
Superiority of Multi-Head Attention in In-Context Linear Regression
by: Cui, Yingqian, et al.
Published: (2024)
by: Cui, Yingqian, et al.
Published: (2024)
On the Mechanism and Dynamics of Modular Addition: Fourier Features, Lottery Ticket, and Grokking
by: He, Jianliang, et al.
Published: (2026)
by: He, Jianliang, et al.
Published: (2026)
From Words to Actions: Unveiling the Theoretical Underpinnings of LLM-Driven Autonomous Systems
by: He, Jianliang, et al.
Published: (2024)
by: He, Jianliang, et al.
Published: (2024)
Active-Dormant Attention Heads: Mechanistically Demystifying Extreme-Token Phenomena in LLMs
by: Guo, Tianyu, et al.
Published: (2024)
by: Guo, Tianyu, et al.
Published: (2024)
Unveiling Induction Heads: Provable Training Dynamics and Feature Learning in Transformers
by: Chen, Siyu, et al.
Published: (2024)
by: Chen, Siyu, et al.
Published: (2024)
Beyond Linear Attention: Softmax Transformers Implement In-Context Reinforcement Learning
by: Xie, Zixuan, et al.
Published: (2026)
by: Xie, Zixuan, et al.
Published: (2026)
Local Linear Attention: An Optimal Interpolation of Linear and Softmax Attention For Test-Time Regression
by: Zuo, Yifei, et al.
Published: (2025)
by: Zuo, Yifei, et al.
Published: (2025)
How Transformers Utilize Multi-Head Attention in In-Context Learning? A Case Study on Sparse Linear Regression
by: Chen, Xingwu, et al.
Published: (2024)
by: Chen, Xingwu, et al.
Published: (2024)
Sample-efficient Learning of Infinite-horizon Average-reward MDPs with General Function Approximation
by: He, Jianliang, et al.
Published: (2024)
by: He, Jianliang, et al.
Published: (2024)
Training Dynamics of In-Context Learning in Linear Attention
by: Zhang, Yedi, et al.
Published: (2025)
by: Zhang, Yedi, et al.
Published: (2025)
Why Softmax Attention Outperforms Linear Attention
by: Deng, Yichuan, et al.
Published: (2023)
by: Deng, Yichuan, et al.
Published: (2023)
Training Dynamics of Softmax Self-Attention: Fast Global Convergence via Preconditioning
by: Goel, Gautam, et al.
Published: (2026)
by: Goel, Gautam, et al.
Published: (2026)
Implicit Regularization of Gradient Flow on One-Layer Softmax Attention
by: Sheen, Heejune, et al.
Published: (2024)
by: Sheen, Heejune, et al.
Published: (2024)
Degrees of Freedom for Linear Attention: Distilling Softmax Attention with Optimal Feature Efficiency
by: Nishikawa, Naoki, et al.
Published: (2025)
by: Nishikawa, Naoki, et al.
Published: (2025)
MetaLA: Unified Optimal Linear Approximation to Softmax Attention Map
by: Chou, Yuhong, et al.
Published: (2024)
by: Chou, Yuhong, et al.
Published: (2024)
The Hedgehog & the Porcupine: Expressive Linear Attentions with Softmax Mimicry
by: Zhang, Michael, et al.
Published: (2024)
by: Zhang, Michael, et al.
Published: (2024)
Demystifying the Slash Pattern in Attention: The Role of RoPE
by: Cheng, Yuan, et al.
Published: (2026)
by: Cheng, Yuan, et al.
Published: (2026)
Universal Approximation with Softmax Attention
by: Hu, Jerry Yao-Chieh, et al.
Published: (2025)
by: Hu, Jerry Yao-Chieh, et al.
Published: (2025)
Softmax as Linear Attention in the Large-Prompt Regime: a Measure-based Perspective
by: Boursier, Etienne, et al.
Published: (2025)
by: Boursier, Etienne, et al.
Published: (2025)
Softmax Linear Attention: Reclaiming Global Competition
by: Xu, Mingwei, et al.
Published: (2026)
by: Xu, Mingwei, et al.
Published: (2026)
Value-Based Deep Multi-Agent Reinforcement Learning with Dynamic Sparse Training
by: Hu, Pihe, et al.
Published: (2024)
by: Hu, Pihe, et al.
Published: (2024)
Statistical Advantage of Softmax Attention: Insights from Single-Location Regression
by: Duranthon, O., et al.
Published: (2025)
by: Duranthon, O., et al.
Published: (2025)
Mechanistic Data Attribution: Tracing the Training Origins of Interpretable LLM Units
by: Chen, Jianhui, et al.
Published: (2026)
by: Chen, Jianhui, et al.
Published: (2026)
Dynamics of Transient Structure in In-Context Linear Regression Transformers
by: Carroll, Liam, et al.
Published: (2025)
by: Carroll, Liam, et al.
Published: (2025)
In-Context Learning with Transformers: Softmax Attention Adapts to Function Lipschitzness
by: Collins, Liam, et al.
Published: (2024)
by: Collins, Liam, et al.
Published: (2024)
Demystifying Linear MDPs and Novel Dynamics Aggregation Framework
by: Lee, Joongkyu, et al.
Published: (2024)
by: Lee, Joongkyu, et al.
Published: (2024)
Trained Mamba Emulates Online Gradient Descent in In-Context Linear Regression
by: Jiang, Jiarui, et al.
Published: (2025)
by: Jiang, Jiarui, et al.
Published: (2025)
Shortcut to Nowhere: Demystifying Deep Spurious Regression
by: Xu, Guanrong, et al.
Published: (2026)
by: Xu, Guanrong, et al.
Published: (2026)
Scalable-Softmax Is Superior for Attention
by: Nakanishi, Ken M.
Published: (2025)
by: Nakanishi, Ken M.
Published: (2025)
Mechanistic Interpretability of Fine-Tuned Vision Transformers on Distorted Images: Decoding Attention Head Behavior for Transparent and Trustworthy AI
by: Bahador, Nooshin
Published: (2025)
by: Bahador, Nooshin
Published: (2025)
Multi-Head Low-Rank Attention
by: Liu, Songtao, et al.
Published: (2026)
by: Liu, Songtao, et al.
Published: (2026)
Forgetting Transformer: Softmax Attention with a Forget Gate
by: Lin, Zhixuan, et al.
Published: (2025)
by: Lin, Zhixuan, et al.
Published: (2025)
Model Collapse Demystified: The Case of Regression
by: Dohmatob, Elvis, et al.
Published: (2024)
by: Dohmatob, Elvis, et al.
Published: (2024)
Softpick: No Attention Sink, No Massive Activations with Rectified Softmax
by: Zuhri, Zayd M. K., et al.
Published: (2025)
by: Zuhri, Zayd M. K., et al.
Published: (2025)
Are Linear Regression Models White Box and Interpretable?
by: Salih, Ahmed M, et al.
Published: (2024)
by: Salih, Ahmed M, et al.
Published: (2024)
On the Invariants of Softmax Attention
by: Lee, Wonsuk
Published: (2026)
by: Lee, Wonsuk
Published: (2026)
In-context Learning for Mixture of Linear Regressions: Existence, Generalization and Training Dynamics
by: Jin, Yanhao, et al.
Published: (2024)
by: Jin, Yanhao, et al.
Published: (2024)
Adaptive Sampled Softmax with Inverted Multi-Index: Methods, Theory and Applications
by: Chen, Jin, et al.
Published: (2025)
by: Chen, Jin, et al.
Published: (2025)
The Non-Linear Representation Dilemma: Is Causal Abstraction Enough for Mechanistic Interpretability?
by: Sutter, Denis, et al.
Published: (2025)
by: Sutter, Denis, et al.
Published: (2025)
Similar Items
-
Training Dynamics of Multi-Head Softmax Attention for In-Context Learning: Emergence, Convergence, and Optimality
by: Chen, Siyu, et al.
Published: (2024) -
Superiority of Multi-Head Attention in In-Context Linear Regression
by: Cui, Yingqian, et al.
Published: (2024) -
On the Mechanism and Dynamics of Modular Addition: Fourier Features, Lottery Ticket, and Grokking
by: He, Jianliang, et al.
Published: (2026) -
From Words to Actions: Unveiling the Theoretical Underpinnings of LLM-Driven Autonomous Systems
by: He, Jianliang, et al.
Published: (2024) -
Active-Dormant Attention Heads: Mechanistically Demystifying Extreme-Token Phenomena in LLMs
by: Guo, Tianyu, et al.
Published: (2024)