# Relative Importance in Sentence Processing

arXiv: https://arxiv.org/abs/2106.03471 · HTML: https://mlsanity.com/paper/2106.03471/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 56 references that are in the map.
- First posted on arXiv 7 Jun 2021.

## Builds on (most cited first, 27 of 27)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Neural Machine Translation by Jointly Learning to Align and Translate](https://mlsanity.com/paper/1409.0473/) (2014)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://mlsanity.com/paper/1909.11942/) (2019)
- [Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps](https://mlsanity.com/paper/1312.6034/) (2013)
- [Show, Attend and Tell: Neural Image Caption Generation with Visual Attention](https://mlsanity.com/paper/1502.03044/) (2015)
- [Attention is not Explanation](https://mlsanity.com/paper/1902.10186/) (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)
- [Is Attention Interpretable?](https://mlsanity.com/paper/1906.03731/) (2019)
- [A Multiscale Visualization of Attention in the Transformer Model](https://mlsanity.com/paper/1906.05714/) (2019)
- [Understanding Neural Networks through Representation Erasure](https://mlsanity.com/paper/1612.08220/) (2016)
- [Colorless green recurrent networks dream hierarchically](https://mlsanity.com/paper/1803.11138/) (2018)
- [The elephant in the interpretability room: Why use attention as explanation when we have saliency methods?](https://mlsanity.com/paper/2010.05607/) (2020)
- [Ordered Neurons: Integrating Tree Structures into Recurrent Neural Networks](https://mlsanity.com/paper/1810.09536/) (2018)
- [ZuCo 2.0: A Dataset of Physiological Recordings During Natural Reading and Annotation](https://mlsanity.com/paper/1912.00903/) (2019)
- [Improving Natural Language Processing Tasks with Human Gaze-Guided Neural Attention](https://mlsanity.com/paper/2010.07891/) (2020)
- [Evaluating Saliency Methods for Neural Language Models](https://mlsanity.com/paper/2104.05824/) (2021)
- [Interpreting Attention Models with Human Visual Attention in Machine Reading Comprehension](https://mlsanity.com/paper/2010.06396/) (2020)
- [Multilingual Language Models Predict Human Reading Behavior](https://mlsanity.com/paper/2104.05433/) (2021)
- [Saliency-driven Word Alignment Interpretation for Neural Machine Translation](https://mlsanity.com/paper/1906.10282/) (2019)
- [Bridging Information-Seeking Human Gaze and Machine Reading Comprehension](https://mlsanity.com/paper/2009.14780/) (2020)
- [Generating Image Descriptions via Sequential Cross-Modal Alignment Guided by Human Gaze](https://mlsanity.com/paper/2011.04592/) (2020)
- [Multimodal Grounding for Language Processing](https://mlsanity.com/paper/1806.06371/) (2018)
- [Semantic Drift in Multilingual Representations](https://mlsanity.com/paper/1904.10820/) (2019)
- [A Discriminative Model for Identifying Readers and Assessing Text Comprehension from Eye Movements](https://mlsanity.com/paper/1809.08031/) (2018)
- [How much complexity does an RNN architecture need to learn syntax-sensitive dependencies?](https://mlsanity.com/paper/2005.08199/) (2020)

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

- [Why Does Surprisal From Larger Transformer-Based Language Models Provide a Poorer Fit to Human Reading Times?](https://mlsanity.com/paper/2212.12131/) (2022)
- [XAI for Transformers: Better Explanations through Conservative Propagation](https://mlsanity.com/paper/2202.07304/) (2022)
- [Do Transformer Models Show Similar Attention Patterns to Task-Specific Human Gaze?](https://mlsanity.com/paper/2205.10226/) (2022)
- [Predicting Fine-Tuning Performance with Probing](https://mlsanity.com/paper/2210.07352/) (2022)
- [Seeing Eye to AI: Human Alignment via Gaze-Based Response Rewards for Large Language Models](https://mlsanity.com/paper/2410.01532/) (2024)
- [The Copenhagen Corpus of Eye Tracking Recordings from Natural Reading of Danish Texts](https://mlsanity.com/paper/2204.13311/) (2022)
- [Roles of Scaling and Instruction Tuning in Language Perception: Model vs. Human Attention](https://mlsanity.com/paper/2310.19084/) (2023)
- [WebQAmGaze: A Multilingual Webcam Eye-Tracking-While-Reading Dataset](https://mlsanity.com/paper/2303.17876/) (2023)
- [Progressive Inference: Explaining Decoder-Only Sequence Classification Models Using Intermediate Predictions](https://mlsanity.com/paper/2406.02625/) (2024)
- [Reading Task Classification Using EEG and Eye-Tracking Data](https://mlsanity.com/paper/2112.06310/) (2021)
- [A Multilingual Perspective Towards the Evaluation of Attribution Methods in Natural Language Inference](https://mlsanity.com/paper/2204.05428/) (2022)
- [EMTeC: A Corpus of Eye Movements on Machine-Generated Texts](https://mlsanity.com/paper/2408.04289/) (2024)
- [Explainable Compliance Detection with Multi-Hop Natural Language Inference on Assurance Case Structure](https://mlsanity.com/paper/2506.08713/) (2025)
- [Every word counts: A multilingual analysis of individual human alignment with model attention](https://mlsanity.com/paper/2210.04963/) (2022)
- [OASST-ETC Dataset: Alignment Signals from Eye-tracking Analysis of LLM Responses](https://mlsanity.com/paper/2503.10927/) (2025)
- [B-Cos Aligned Transformers Learn Human-Interpretable Features](https://mlsanity.com/paper/2401.08868/) (2024)
- [Integrating Cognitive Processing Signals into Language Models: A Review of Advances, Applications and Future Directions](https://mlsanity.com/paper/2504.06843/) (2025)
- [Measuring Alignment With Reader Highlights Net of Position and Length](https://mlsanity.com/paper/2607.27739/) (2026)
- [Explainable AI: Context-Aware Layer-Wise Integrated Gradients for Explaining Transformer Models](https://mlsanity.com/paper/2602.16608/) (2026)
- [NU HLT at CMCL 2022 Shared Task: Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space](https://mlsanity.com/paper/2202.10855/) (2022)
- [Text Difficulty Study: Do machines behave the same as humans regarding text difficulty?](https://mlsanity.com/paper/2208.14509/) (2022)
- [Evaluating Webcam-based Gaze Data as an Alternative for Human Rationale Annotations](https://mlsanity.com/paper/2402.19133/) (2024)

## Abstract (arXiv)

Determining the relative importance of the elements in a sentence is a key factor for effortless natural language understanding. For human language processing, we can approximate patterns of relative importance by measuring reading fixations using eye-tracking technology. In neural language models, gradient-based saliency methods indicate the relative importance of a token for the target objective. In this work, we compare patterns of relative importance in English language processing by humans and models and analyze the underlying linguistic patterns. We find that human processing patterns in English correlate strongly with saliency-based importance in language models and not with attention-based importance. Our results indicate that saliency could be a cognitively more plausible metric for interpreting neural language models. The code is available on GitHub: https://github.com/beinborn/relative_importance