# Learning to Faithfully Rationalize by Construction

arXiv: https://arxiv.org/abs/2005.00115 · HTML: https://mlsanity.com/paper/2005.00115/

## Key facts

- Cited by 113 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 30 Apr 2020.

## Built on by (most cited first, 30 of 113)

- [Towards Faithful Model Explanation in NLP: A Survey](https://mlsanity.com/paper/2209.11326/) (2022)
- [Explanations from Large Language Models Make Small Reasoners Better](https://mlsanity.com/paper/2210.06726/) (2022)
- [Do Feature Attribution Methods Correctly Attribute Features?](https://mlsanity.com/paper/2104.14403/) (2021)
- [Paragraph-level Rationale Extraction through Regularization: A case study on European Court of Human Rights Cases](https://mlsanity.com/paper/2103.13084/) (2021)
- [PINTO: Faithful Language Reasoning Using Prompt-Generated Rationales](https://mlsanity.com/paper/2211.01562/) (2022)
- [ProoFVer: Natural Logic Theorem Proving for Fact Verification](https://mlsanity.com/paper/2108.11357/) (2021)
- [SelfExplain: A Self-Explaining Architecture for Neural Text Classifiers](https://mlsanity.com/paper/2103.12279/) (2021)
- [Grad-SAM: Explaining Transformers via Gradient Self-Attention Maps](https://mlsanity.com/paper/2204.11073/) (2022)
- [Can Rationalization Improve Robustness?](https://mlsanity.com/paper/2204.11790/) (2022)
- [Going Beyond XAI: A Systematic Survey for Explanation-Guided Learning](https://mlsanity.com/paper/2212.03954/) (2022)
- [UNIREX: A Unified Learning Framework for Language Model Rationale Extraction](https://mlsanity.com/paper/2112.08802/) (2021)
- [DecompX: Explaining Transformers Decisions by Propagating Token Decomposition](https://mlsanity.com/paper/2306.02873/) (2023)
- [Improving the Faithfulness of Attention-based Explanations with Task-specific Information for Text Classification](https://mlsanity.com/paper/2105.02657/) (2021)
- [The Eval4NLP Shared Task on Explainable Quality Estimation: Overview and Results](https://mlsanity.com/paper/2110.04392/) (2021)
- [Rationales for Sequential Predictions](https://mlsanity.com/paper/2109.06387/) (2021)
- [What to Learn, and How: Toward Effective Learning from Rationales](https://mlsanity.com/paper/2112.00071/) (2021)
- [On Behalf of the Stakeholders: Trends in NLP Model Interpretability in the Era of LLMs](https://mlsanity.com/paper/2407.19200/) (2024)
- [Do Large Language Models Perform the Way People Expect? Measuring the Human Generalization Function](https://mlsanity.com/paper/2406.01382/) (2024)
- [Detect and Perturb: Neutral Rewriting of Biased and Sensitive Text via Gradient-based Decoding](https://mlsanity.com/paper/2109.11708/) (2021)
- [Translation Error Detection as Rationale Extraction](https://mlsanity.com/paper/2108.12197/) (2021)
- [REV: Information-Theoretic Evaluation of Free-Text Rationales](https://mlsanity.com/paper/2210.04982/) (2022)
- [Self-Interpretable Model with TransformationEquivariant Interpretation](https://mlsanity.com/paper/2111.04927/) (2021)
- [An Empirical Study on Explanations in Out-of-Domain Settings](https://mlsanity.com/paper/2203.00056/) (2022)
- [ProtoTEx: Explaining Model Decisions with Prototype Tensors](https://mlsanity.com/paper/2204.05426/) (2022)
- [D-Separation for Causal Self-Explanation](https://mlsanity.com/paper/2309.13391/) (2023)
- [Tailoring Self-Rationalizers with Multi-Reward Distillation](https://mlsanity.com/paper/2311.02805/) (2023)
- [Rationalization through Concepts](https://mlsanity.com/paper/2105.04837/) (2021)
- [SPECTRA: Sparse Structured Text Rationalization](https://mlsanity.com/paper/2109.04552/) (2021)
- [Interpreting Deep Learning Models in Natural Language Processing: A Review](https://mlsanity.com/paper/2110.10470/) (2021)
- [AdapLeR: Speeding up Inference by Adaptive Length Reduction](https://mlsanity.com/paper/2203.08991/) (2022)

## Abstract (arXiv)

In many settings it is important for one to be able to understand why a model made a particular prediction. In NLP this often entails extracting snippets of an input text `responsible for' corresponding model output; when such a snippet comprises tokens that indeed informed the model's prediction, it is a faithful explanation. In some settings, faithfulness may be critical to ensure transparency. Lei et al. (2016) proposed a model to produce faithful rationales for neural text classification by defining independent snippet extraction and prediction modules. However, the discrete selection over input tokens performed by this method complicates training, leading to high variance and requiring careful hyperparameter tuning. We propose a simpler variant of this approach that provides faithful explanations by construction. In our scheme, named FRESH, arbitrary feature importance scores (e.g., gradients from a trained model) are used to induce binary labels over token inputs, which an extractor can be trained to predict. An independent classifier module is then trained exclusively on snippets provided by the extractor; these snippets thus constitute faithful explanations, even if the classifier is arbitrarily complex. In both automatic and manual evaluations we find that variants of this simple framework yield predictive performance superior to `end-to-end' approaches, while being more general and easier to train. Code is available at https://github.com/successar/FRESH