# Incorporating Attribution Importance for Improving Faithfulness Metrics

arXiv: https://arxiv.org/abs/2305.10496 · HTML: https://mlsanity.com/paper/2305.10496/

## Key facts

- Cited by 14 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 36 of its 48 references that are in the map.
- First posted on arXiv 17 May 2023.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 30 of 36)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [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)
- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [Visualizing and Understanding Convolutional Networks](https://mlsanity.com/paper/1311.2901/) (2013)
- [Learning Important Features Through Propagating Activation Differences](https://mlsanity.com/paper/1704.02685/) (2017)
- [SciBERT: A Pretrained Language Model for Scientific Text](https://mlsanity.com/paper/1903.10676/) (2019)
- [Attention is not Explanation](https://mlsanity.com/paper/1902.10186/) (2019)
- [Towards Faithfully Interpretable NLP Systems: How should we define and evaluate faithfulness?](https://mlsanity.com/paper/2004.03685/) (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)
- [On the Robustness of Interpretability Methods](https://mlsanity.com/paper/1806.08049/) (2018)
- [One Explanation Does Not Fit All: A Toolkit and Taxonomy of AI Explainability Techniques](https://mlsanity.com/paper/1909.03012/) (2019)
- [A Diagnostic Study of Explainability Techniques for Text Classification](https://mlsanity.com/paper/2009.13295/) (2020)
- [Towards Faithful Model Explanation in NLP: A Survey](https://mlsanity.com/paper/2209.11326/) (2022)
- [The elephant in the interpretability room: Why use attention as explanation when we have saliency methods?](https://mlsanity.com/paper/2010.05607/) (2020)
- [Learning to Faithfully Rationalize by Construction](https://mlsanity.com/paper/2005.00115/) (2020)
- [Paragraph-level Rationale Extraction through Regularization: A case study on European Court of Human Rights Cases](https://mlsanity.com/paper/2103.13084/) (2021)
- [Inferring Which Medical Treatments Work from Reports of Clinical Trials](https://mlsanity.com/paper/1904.01606/) (2019)
- [Investigating the influence of noise and distractors on the interpretation of neural networks](https://mlsanity.com/paper/1611.07270/) (2016)
- ["What is Relevant in a Text Document?": An Interpretable Machine Learning Approach](https://mlsanity.com/paper/1612.07843/) (2016)
- [A Comparative Study of Faithfulness Metrics for Model Interpretability Methods](https://mlsanity.com/paper/2204.05514/) (2022)
- [Evaluating and Characterizing Human Rationales](https://mlsanity.com/paper/2010.04736/) (2020)
- [On the Sensitivity and Stability of Model Interpretations in NLP](https://mlsanity.com/paper/2104.08782/) (2021)
- [Perturbing Inputs for Fragile Interpretations in Deep Natural Language Processing](https://mlsanity.com/paper/2108.04990/) (2021)
- [Improving the Faithfulness of Attention-based Explanations with Task-specific Information for Text Classification](https://mlsanity.com/paper/2105.02657/) (2021)
- [Translation Error Detection as Rationale Extraction](https://mlsanity.com/paper/2108.12197/) (2021)
- [An Empirical Study on Explanations in Out-of-Domain Settings](https://mlsanity.com/paper/2203.00056/) (2022)
- [Connecting Attributions and QA Model Behavior on Realistic Counterfactuals](https://mlsanity.com/paper/2104.04515/) (2021)
- [Staying True to Your Word: (How) Can Attention Become Explanation?](https://mlsanity.com/paper/2005.09379/) (2020)

## Built on by (most cited first, 14 of 14)

- [Investigating Hallucinations in Pruned Large Language Models for Abstractive Summarization](https://mlsanity.com/paper/2311.09335/) (2023)
- [Why and How LLMs Hallucinate: Connecting the Dots with Subsequence Associations](https://mlsanity.com/paper/2504.12691/) (2025)
- [Latent Concept-based Explanation of NLP Models](https://mlsanity.com/paper/2404.12545/) (2024)
- [ReAGent: A Model-agnostic Feature Attribution Method for Generative Language Models](https://mlsanity.com/paper/2402.00794/) (2024)
- [Normalized AOPC: Fixing Misleading Faithfulness Metrics for Feature Attribution Explainability](https://mlsanity.com/paper/2408.08137/) (2024)
- [Comparing Explanation Faithfulness between Multilingual and Monolingual Fine-tuned Language Models](https://mlsanity.com/paper/2403.12809/) (2024)
- [Noiser: Bounded Input Perturbations for Attributing Large Language Models](https://mlsanity.com/paper/2504.02911/) (2025)
- [Gender Bias in Explainability: Investigating Performance Disparity in Post-hoc Methods](https://mlsanity.com/paper/2505.01198/) (2025)
- [DLBacktrace: A Model Agnostic Explainability for any Deep Learning Models](https://mlsanity.com/paper/2411.12643/) (2024)
- [EvalxNLP: A Framework for Benchmarking Post-Hoc Explainability Methods on NLP Models](https://mlsanity.com/paper/2505.01238/) (2025)
- [Cross-Layer Discrete Concept Discovery for Interpreting Language Models](https://mlsanity.com/paper/2506.20040/) (2025)
- [Hessian-Enhanced Token Attribution (HETA): Interpreting Autoregressive LLMs](https://mlsanity.com/paper/2604.13258/) (2026)
- [Support-Contra Asymmetry in LLM Explanations](https://mlsanity.com/paper/2510.21884/) (2025)
- [ExU: AI Models for Examining Multilingual Disinformation Narratives and Understanding their Spread](https://mlsanity.com/paper/2406.15443/) (2024)

## Abstract (arXiv)

Feature attribution methods (FAs) are popular approaches for providing insights into the model reasoning process of making predictions. The more faithful a FA is, the more accurately it reflects which parts of the input are more important for the prediction. Widely used faithfulness metrics, such as sufficiency and comprehensiveness use a hard erasure criterion, i.e. entirely removing or retaining the top most important tokens ranked by a given FA and observing the changes in predictive likelihood. However, this hard criterion ignores the importance of each individual token, treating them all equally for computing sufficiency and comprehensiveness. In this paper, we propose a simple yet effective soft erasure criterion. Instead of entirely removing or retaining tokens from the input, we randomly mask parts of the token vector representations proportionately to their FA importance. Extensive experiments across various natural language processing tasks and different FAs show that our soft-sufficiency and soft-comprehensiveness metrics consistently prefer more faithful explanations compared to hard sufficiency and comprehensiveness. Our code: https://github.com/casszhao/SoftFaith