# Measuring the Mixing of Contextual Information in the Transformer

arXiv: https://arxiv.org/abs/2203.04212 · HTML: https://mlsanity.com/paper/2203.04212/

## Key facts

- Cited by 65 papers in the map (data through 14 Oct 2026).
- Builds on 44 of its 58 references that are in the map.
- First posted on arXiv 8 Mar 2022.

## Builds on (most cited first, 30 of 44)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- ["Why Should I Trust You?": Explaining the Predictions of Any Classifier](https://mlsanity.com/paper/1602.04938/) (2016)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps](https://mlsanity.com/paper/1312.6034/) (2013)
- [Character-level Convolutional Networks for Text Classification](https://mlsanity.com/paper/1509.01626/) (2015)
- [Visualizing and Understanding Convolutional Networks](https://mlsanity.com/paper/1311.2901/) (2013)
- [What Does BERT Look At? An Analysis of BERT's Attention](https://mlsanity.com/paper/1906.04341/) (2019)
- [Attention is not Explanation](https://mlsanity.com/paper/1902.10186/) (2019)
- [How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings](https://mlsanity.com/paper/1909.00512/) (2019)
- [Quantifying Attention Flow in Transformers](https://mlsanity.com/paper/2005.00928/) (2020)
- [Attention is not not Explanation](https://mlsanity.com/paper/1908.04626/) (2019)
- [Transformer Interpretability Beyond Attention Visualization](https://mlsanity.com/paper/2012.09838/) (2020)
- [Captum: A unified and generic model interpretability library for PyTorch](https://mlsanity.com/paper/2009.07896/) (2020)
- [ERASER: A Benchmark to Evaluate Rationalized NLP Models](https://mlsanity.com/paper/1911.03429/) (2019)
- [Is Attention Interpretable?](https://mlsanity.com/paper/1906.03731/) (2019)
- [Assessing the Ability of LSTMs to Learn Syntax-Sensitive Dependencies](https://mlsanity.com/paper/1611.01368/) (2016)
- [Generic Attention-model Explainability for Interpreting Bi-Modal and Encoder-Decoder Transformers](https://mlsanity.com/paper/2103.15679/) (2021)
- [Revealing the Dark Secrets of BERT](https://mlsanity.com/paper/1908.08593/) (2019)
- [Not Just a Black Box: Learning Important Features Through Propagating Activation Differences](https://mlsanity.com/paper/1605.01713/) (2016)
- [Analyzing the Structure of Attention in a Transformer Language Model](https://mlsanity.com/paper/1906.04284/) (2019)
- [Visualizing and Understanding Neural Models in NLP](https://mlsanity.com/paper/1506.01066/) (2015)
- [Assessing BERT's Syntactic Abilities](https://mlsanity.com/paper/1901.05287/) (2019)
- [A Diagnostic Study of Explainability Techniques for Text Classification](https://mlsanity.com/paper/2009.13295/) (2020)
- [Post-hoc Interpretability for Neural NLP: A Survey](https://mlsanity.com/paper/2108.04840/) (2021)
- [All Bark and No Bite: Rogue Dimensions in Transformer Language Models Obscure Representational Quality](https://mlsanity.com/paper/2109.04404/) (2021)

## Built on by (most cited first, 30 of 65)

- [Information Flow Routes: Automatically Interpreting Language Models at Scale](https://mlsanity.com/paper/2403.00824/) (2024)
- [From Understanding to Utilization: A Survey on Explainability for Large Language Models](https://mlsanity.com/paper/2401.12874/) (2024)
- [Towards Opening the Black Box of Neural Machine Translation: Source and Target Interpretations of the Transformer](https://mlsanity.com/paper/2205.11631/) (2022)
- [Quantifying Context Mixing in Transformers](https://mlsanity.com/paper/2301.12971/) (2023)
- [Explaining How Transformers Use Context to Build Predictions](https://mlsanity.com/paper/2305.12535/) (2023)
- [Analyzing Feed-Forward Blocks in Transformers through the Lens of Attention Maps](https://mlsanity.com/paper/2302.00456/) (2023)
- [DecompX: Explaining Transformers Decisions by Propagating Token Decomposition](https://mlsanity.com/paper/2306.02873/) (2023)
- [InternalInspector I2: Robust Confidence Estimation in LLMs through Internal States](https://mlsanity.com/paper/2406.12053/) (2024)
- [Toxicity in Multilingual Machine Translation at Scale](https://mlsanity.com/paper/2210.03070/) (2022)
- [ByGPT5: End-to-End Style-conditioned Poetry Generation with Token-free Language Models](https://mlsanity.com/paper/2212.10474/) (2022)
- [LM Transparency Tool: Interactive Tool for Analyzing Transformer Language Models](https://mlsanity.com/paper/2404.07004/) (2024)
- [Homophone Disambiguation Reveals Patterns of Context Mixing in Speech Transformers](https://mlsanity.com/paper/2310.09925/) (2023)
- [Evaluating Human Alignment and Model Faithfulness of LLM Rationale](https://mlsanity.com/paper/2407.00219/) (2024)
- [A Close Look at Decomposition-based XAI-Methods for Transformer Language Models](https://mlsanity.com/paper/2502.15886/) (2025)
- [The Atlas of In-Context Learning: How Attention Heads Shape In-Context Retrieval Augmentation](https://mlsanity.com/paper/2505.15807/) (2025)
- [Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety](https://mlsanity.com/paper/2506.05451/) (2025)
- [Feature Interactions Reveal Linguistic Structure in Language Models](https://mlsanity.com/paper/2306.12181/) (2023)
- [Better Explain Transformers by Illuminating Important Information](https://mlsanity.com/paper/2401.09972/) (2024)
- [Quantifying the Plausibility of Context Reliance in Neural Machine Translation](https://mlsanity.com/paper/2310.01188/) (2023)
- [Roles of Scaling and Instruction Tuning in Language Perception: Model vs. Human Attention](https://mlsanity.com/paper/2310.19084/) (2023)
- [Are We Paying Attention to Her? Investigating Gender Disambiguation and Attention in Machine Translation](https://mlsanity.com/paper/2505.08546/) (2025)
- [Towards Long-Horizon Interpretability: Efficient and Faithful Multi-Token Attribution for Reasoning LLMs](https://mlsanity.com/paper/2602.01914/) (2026)
- [Unveiling and Manipulating Prompt Influence in Large Language Models](https://mlsanity.com/paper/2405.11891/) (2024)
- [Context-Aware or Context-Insensitive? Assessing LLMs' Performance in Document-Level Translation](https://mlsanity.com/paper/2410.14391/) (2024)
- [Mitigating Hallucinated Translations in Large Language Models with Hallucination-focused Preference Optimization](https://mlsanity.com/paper/2501.17295/) (2025)
- [LaTIM: Measuring Latent Token-to-Token Interactions in Mamba Models](https://mlsanity.com/paper/2502.15612/) (2025)
- [Instruction Anchor: Dissecting the Mechanistic Dynamics of Modality Arbitration](https://mlsanity.com/paper/2602.03677/) (2026)
- [ReX: A Framework for Incorporating Temporal Information in Model-Agnostic Local Explanation Techniques](https://mlsanity.com/paper/2209.03798/) (2022)
- [LibraGrad: Balancing Gradient Flow for Universally Better Vision Transformer Attributions](https://mlsanity.com/paper/2411.16760/) (2024)
- [Geometry-Calibrated Conformal Abstention for Language Models](https://mlsanity.com/paper/2604.27914/) (2026)

## Abstract (arXiv)

The Transformer architecture aggregates input information through the self-attention mechanism, but there is no clear understanding of how this information is mixed across the entire model. Additionally, recent works have demonstrated that attention weights alone are not enough to describe the flow of information. In this paper, we consider the whole attention block -- multi-head attention, residual connection, and layer normalization -- and define a metric to measure token-to-token interactions within each layer. Then, we aggregate layer-wise interpretations to provide input attribution scores for model predictions. Experimentally, we show that our method, ALTI (Aggregation of Layer-wise Token-to-token Interactions), provides more faithful explanations and increased robustness than gradient-based methods.