mlsanityData to 14 Oct 2026

Towards Consistent Natural-Language Explanations via Explanation-Consistency Finetuning

arXiv 2401.13986 ↗First posted 25 Jan 2024cs.CL cs.AI cs.LGcited by 26

Not on the map. Why?

Abstract (arXiv)

Large language models (LLMs) often generate convincing, fluent explanations. However, different from humans, they often generate inconsistent explanations on different inputs. For example, an LLM may generate the explanation "all birds can fly" when answering the question "Can sparrows fly?" but meanwhile answer "no" to the related question "Can penguins fly?". Explanations should be consistent across related examples so that they allow a human to simulate the LLM's decision process on multiple examples. We propose explanation-consistency finetuning (EC-finetuning), a method that adapts LLMs to generate more consistent natural-language explanations on related examples. EC-finetuning involves finetuning LLMs on synthetic data that is carefully constructed to contain consistent explanations. Across a variety of question-answering datasets in various domains, EC-finetuning yields a 10.0% relative explanation consistency improvement on four finetuning datasets, and generalizes to seven out-of-distribution datasets not seen during finetuning (+4.5% relative). Code is available at https://github.com/yandachen/explanation-consistency-finetuning .

Key facts. Cited by 26 papers in the map (data through 14 Oct 2026). Builds on 27 of its 42 references that are in the map. First posted on arXiv 25 Jan 2024.

Builds on

27 of 42 references.

References · 27

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  3. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,626
  4. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  5. 2018Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challengecited by 4,886
  6. 2021Finetuned Language Models Are Zero-Shot Learnerscited by 3,780
  7. 2020Learning to summarize from human feedbackcited by 2,588
  8. 2019BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questionscited by 2,183
  9. 2018Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answeringcited by 2,164
  10. 2016Concrete Problems in AI Safetycited by 1,642
  11. 2016MS MARCO: A Human Generated MAchine Reading COmprehension Datasetcited by 1,605
  12. 2019PubMedQA: A Dataset for Biomedical Research Question Answeringcited by 1,288
15 more
  1. 2023Quantifying Language Models' Sensitivity to Spurious Features in Prompt Design or: How I learned to start worrying about prompt formattingcited by 817
  2. 2021Did Aristotle Use a Laptop? A Question Answering Benchmark with Implicit Reasoning Strategiescited by 743
  3. 2023Measuring Faithfulness in Chain-of-Thought Reasoningcited by 703
  4. 2020Towards Faithfully Interpretable NLP Systems: How should we define and evaluate faithfulness?cited by 599
  5. 2020Artificial Intelligence, Values and Alignmentcited by 448
  6. 2019Multi-Task Deep Neural Networks for Natural Language Understandingcited by 350
  7. 2016European Union regulations on algorithmic decision-making and a "right to explanation"cited by 344
  8. 2020Evaluating Explainable AI: Which Algorithmic Explanations Help Users Predict Model Behavior?cited by 190
  9. 2023Tell Your Model Where to Attend: Post-hoc Attention Steering for LLMscited by 91
  10. 2023Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanationscited by 85
  11. 2022On the Relation between Sensitivity and Accuracy in In-context Learningcited by 72
  12. 2019An Evaluation of the Human-Interpretability of Explanationcited by 71
  13. 2017The Promise and Peril of Human Evaluation for Model Interpretabilitycited by 52
  14. 2023Benchmarking and Improving Generator-Validator Consistency of Language Modelscited by 44
  15. 2018Medical Exam Question Answering with Large-scale Reading Comprehensioncited by 29

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 8; 2025: 14; 2026: 4.

26 papers.

Sort:

By area:

  1. 2025Reasoning Models Don't Always Say What They Thinkcited by 415
  2. 2024Looking Inward: Language Models Can Learn About Themselves by Introspectioncited by 113
  3. 2024Rethinking Interpretability in the Era of Large Language Modelscited by 111
  4. 2025Are DeepSeek R1 And Other Reasoning Models More Faithful?cited by 72
  5. 2024Bias-Augmented Consistency Training Reduces Biased Reasoning in Chain-of-Thoughtcited by 39
  6. 2025Teaching Models to Verbalize Reward Hacking in Chain-of-Thought Reasoningcited by 21
  7. 2024Model Tells Itself Where to Attend: Faithfulness Meets Automatic Attention Steeringcited by 17
  8. 2025Can We Predict Alignment Before Models Finish Thinking? Towards Monitoring Misaligned Reasoning Modelscited by 17
  9. 2026Counterfactual Simulation Training for Chain-of-Thought Faithfulnesscited by 17
  10. 2024Crafting Interpretable Embeddings by Asking LLMs Questionscited by 11
  11. 2024Self-Training Meets Consistency: Improving LLMs' Reasoning with Consistency-Driven Rationale Evaluationcited by 6
  12. 2025Reasoning-Grounded Natural Language Explanations for Language Modelscited by 5
14 more
  1. 2024Attribute Structuring Improves LLM-Based Evaluation of Clinical Text Summariescited by 4
  2. 2024XForecast: Evaluating Natural Language Explanations for Time Series Forecastingcited by 4
  3. 2025Aligning What LLMs Do and Say: Towards Self-Consistent Explanationscited by 4
  4. 2025A Necessary Step toward Faithfulness: Measuring and Improving Consistency in Free-Text Explanationscited by 3
  5. 2025Explainability of Large Language Models: Opportunities and Challenges toward Generating Trustworthy Explanationscited by 2
  6. 2026RobustExplain: Evaluating Robustness of LLM-Based Explanation Agents for Recommendationcited by 2
  7. 2025CLARity: Reasoning Consistency Alone Can Teach Reinforced Expertscited by 1
  8. 2025Support-Contra Asymmetry in LLM Explanationscited by 1
  9. 2025Can LLM-Generated Textual Explanations Enhance Model Classification Performance? An Empirical Studycited by 0
  10. 2025Harmonizing Diverse Models: A Layer-wise Merging Strategy for Consistent Generationcited by 0
  11. 2025ProfileXAI: User-Adaptive Explainable AIcited by 0
  12. 2025Two Heads are Better than One: Distilling Large Language Model Features Into Small Models with Feature Decomposition and Mixturecited by 0
  13. 2026Automated Testing of LLM-Based Post Hoc Explainers Using Model Checking as an Oraclecited by 0
  14. 2026A Removal Based Approach to Improve LLM Faithfulness at Test-Timecited by 0