mlsanityData to 14 Oct 2026

On Measuring Faithfulness or Self-consistency of Natural Language Explanations

arXiv 2311.07466 ↗First posted 13 Nov 2023cs.CL cs.AI cs.LGcited by 73

On the map: LLM reasoning → Prompting and chain-of-thought reasoning

Abstract (arXiv)

Large language models (LLMs) can explain their predictions through post-hoc or Chain-of-Thought (CoT) explanations. But an LLM could make up reasonably sounding explanations that are unfaithful to its underlying reasoning. Recent work has designed tests that aim to judge the faithfulness of post-hoc or CoT explanations. In this work we argue that these faithfulness tests do not measure faithfulness to the models' inner workings -- but rather their self-consistency at output level. Our contributions are three-fold: i) We clarify the status of faithfulness tests in view of model explainability, characterising them as self-consistency tests instead. This assessment we underline by ii) constructing a Comparative Consistency Bank for self-consistency tests that for the first time compares existing tests on a common suite of 11 open LLMs and 5 tasks -- including iii) our new self-consistency measure CC-SHAP. CC-SHAP is a fine-grained measure (not a test) of LLM self-consistency. It compares how a model's input contributes to the predicted answer and to generating the explanation. Our fine-grained CC-SHAP metric allows us iii) to compare LLM behaviour when making predictions and to analyse the effect of other consistency tests at a deeper level, which takes us one step further towards measuring faithfulness by bringing us closer to the internals of the model than strictly surface output-oriented tests. Our code is available at urlhttps://github.com/Heidelberg-NLP/CC-SHAP

Key facts. Cited by 73 papers in the map (data through 14 Oct 2026). Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning. Builds on 20 of its 57 references that are in the map. First posted on arXiv 13 Nov 2023.

Builds on

20 of 57 references.

Closest ancestors · 4

Same direction, newest first

  1. 2023Measuring Faithfulness in Chain-of-Thought Reasoning17 Jul
  2. 2023Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting7 May
  3. 2023Faithful Chain-of-Thought Reasoning31 Jan
  4. 2022Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them17 Oct

Other references · 16

  1. 2023GPT-4 Technical Reportcited by 18,396
  2. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  3. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  4. 2023Mistral 7Bcited by 3,164
  5. 2018RISE: Randomized Input Sampling for Explanation of Black-box Modelscited by 838
  6. 2023The RefinedWeb Dataset for Falcon LLM: Outperforming Curated Corpora with Web Data, and Web Data Onlycited by 763
  7. 2022Galactica: A Large Language Model for Sciencecited by 744
  8. 2022Selection-Inference: Exploiting Large Language Models for Interpretable Logical Reasoningcited by 334
  9. 2020WT5?! Training Text-to-Text Models to Explain their Predictionscited by 148
  10. 2022Towards Faithful Model Explanation in NLP: A Surveycited by 130
  11. 2023Question Decomposition Improves the Faithfulness of Model-Generated Reasoningcited by 108
  12. 2023Stay on topic with Classifier-Free Guidancecited by 91
4 more
  1. 2023Can Large Language Models Explain Themselves? A Study of LLM-Generated Self-Explanationscited by 90
  2. 2023Faithful Explanations of Black-box NLP Models Using LLM-generated Counterfactualscited by 44
  3. 2022Gender Artifacts in Visual Datasetscited by 38
  4. 2023In-Context Explainers: Harnessing LLMs for Explaining Black Box Modelscited by 9

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 17; 2025: 31; 2026: 25.

73 papers.

Sort:

By area:

  1. 2024Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spacescited by 611
  2. 2025Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscationcited by 384
  3. 2025Chain-of-Thought Reasoning In The Wild Is Not Always Faithfulcited by 170
  4. 2024Rethinking Interpretability in the Era of Large Language Modelscited by 111
  5. 2025Measuring Chain of Thought Faithfulness by Unlearning Reasoning Stepscited by 47
  6. 2024Reasoning Robustness of LLMs to Adversarial Typographical Errorscited by 31
  7. 2024The Probabilities Also Matter: A More Faithful Metric for Faithfulness of Free-Text Explanations in Large Language Modelscited by 29
  8. 2024Chain-of-Thought Unfaithfulness as Disguised Accuracycited by 28
  9. 2024Do Vision & Language Decoders use Images and Text equally? How Self-consistent are their Explanations?cited by 24
  10. 2024Dissociation of Faithful and Unfaithful Reasoning in LLMscited by 23
  11. 2024Towards Better Chain-of-Thought: A Reflection on Effectiveness and Faithfulnesscited by 20
  12. 2025LLMs Don't Know Their Own Decision Boundaries: The Unreliability of Self-Generated Counterfactual Explanationscited by 16
38 more
  1. 2025Consistency in Language Models: Current Landscape, Challenges, and Future Directionscited by 14
  2. 2026A Positive Case for Faithfulness: LLM Self-Explanations Help Predict Model Behaviorcited by 13
  3. 2025A Causal Lens for Evaluating Faithfulness Metricscited by 12
  4. 2025A Closer Look at Bias and Chain-of-Thought Faithfulness of Large (Vision) Language Modelscited by 12
  5. 2024Multi-Level Explanations for Generative Language Modelscited by 11
  6. 2026RFEval: Benchmarking Reasoning Faithfulness under Counterfactual Reasoning Intervention in Large Reasoning Modelscited by 11
  7. 2025Self-Critique and Refinement for Faithful Natural Language Explanationscited by 10
  8. 2024Chain-of-Probe: Examining the Necessity and Accuracy of CoT Step-by-Stepcited by 8
  9. 2024GraphNarrator: Generating Textual Explanations for Graph Neural Networkscited by 7
  10. 2025Faithful-First Reasoning, Planning, and Acting for Multimodal LLMscited by 7
  11. 2025Translate Policy to Language: Flow Matching Generated Rewards for LLM Explanationscited by 6
  12. 2025Vision-and-Language Training Helps Deploy Taxonomic Knowledge but Does Not Fundamentally Alter Itcited by 6
  13. 2025From Reasoning to Answer: Empirical, Attention-Based and Mechanistic Insights into Distilled DeepSeek R1 Modelscited by 6
  14. 2025Beyond Correctness: Rewarding Faithful Reasoning in Retrieval-Augmented Generationcited by 6
  15. 2026Breaking the Chain: A Causal Analysis of LLM Faithfulness to Intermediate Structurescited by 6
  16. 2024Inference to the Best Explanation in Large Language Modelscited by 5
  17. 2026Can Large Language Models Still Explain Themselves? Investigating the Impact of Quantization on Self-Explanationscited by 5
  18. 2025Aligning What LLMs Do and Say: Towards Self-Consistent Explanationscited by 4
  19. 2026LLMs Can't Play Hangman: On the Necessity of a Private Working Memory for Language Agentscited by 4
  20. 2026Measuring Faithfulness Depends on How You Measure: Classifier Sensitivity in LLM Chain-of-Thought Evaluationcited by 4
  21. 2024Watch Your Steps: Observable and Modular Chains of Thoughtcited by 3
  22. 2024Graph-Guided Textual Explanation Generation Frameworkcited by 3
  23. 2025Quantifying Uncertainty in Natural Language Explanations of Large Language Models for Question Answeringcited by 3
  24. 2025Interpreting Language Models Through Concept Descriptions: A Surveycited by 3
  25. 2025Reasoning Path and Latent State Analysis for Multi-view Visual Spatial Reasoning: A Cognitive Science Perspectivecited by 3
  26. 2026Faithfulness Metrics Don't Measure Faithfulness: A Meta-Evaluation with Ground Truthcited by 3
  27. 2024Clinical information extraction for Low-resource languages with Few-shot learning using Pre-trained language models and Promptingcited by 2
  28. 2025Semi-structured LLM Reasoners Can Be Rigorously Auditedcited by 2
  29. 2025Reranking-based Generation for Unbiased Perspective Summarizationcited by 2
  30. 2025Investigating Modality Contribution in Audio LLMs for Musiccited by 2
  31. 2025Improving Consistency in Retrieval-Augmented Systems with Group Similarity Rewardscited by 2
  32. 2025Explainability of Large Language Models: Opportunities and Challenges toward Generating Trustworthy Explanationscited by 2
  33. 2025Mapping Faithful Reasoning in Language Modelscited by 2
  34. 2025T-FIX: Text-Based Explanations with Features Interpretable to eXpertscited by 2
  35. 2024A Systematic Comparison between Extractive Self-Explanations and Human Rationales in Text Classificationcited by 1
  36. 2024Anchored Alignment for Self-Explanations Enhancementcited by 1
  37. 2025Evaluating Large Language Models for Phishing Detection, Self-Consistency, Faithfulness, and Explainabilitycited by 1
  38. 2025Robust Native Language Identification through Agentic Decompositioncited by 1

50 most cited of 73 All 73 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2024Faithfulness vs. Plausibility: On the (Un)Reliability of Explanations from Large Language Models
  2. 2024Are self-explanations from Large Language Models faithful?
  3. 2024Towards Faithful Natural Language Explanations: A Study Using Activation Patching in Large Language Models
  4. 2024Internal Consistency and Self-Feedback in Large Language Models: A Survey
  5. 2026Does Inference Scaling Improve Reasoning Faithfulness? A Multi-Model Analysis of Self-Consistency Tradeoffs
  6. 2026Making LLMs Say What They Think: Measuring and Improving CoT-Interpretability Alignment
  7. 2026Enhancing Assessment of Self-Consistency in LLM Explanations using Perturbation Strength
  8. 2025A Necessary Step toward Faithfulness: Measuring and Improving Consistency in Free-Text Explanations