Faithfulness Serum: Mitigating the Faithfulness Gap in Textual Explanations of LLM Decisions via Attribution Guidance

arXiv:2604.14325 · cs.CL, cs.AI · Submitted 2026-04-15 · Read on arXiv

cs.CL, cs.AI

Submitted: 2026-04-15

Updated: 2026-04-15

Code: https://github.com/Bar-A-94/FaithfulnessSerum

Terminology

• Benchmark — 1× in this paper · explained in 2 episodes
• Explainability — 1× in this paper · explained in 3 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes

Sources

• Quantifying Attention Flow in Transformers
• AttnLRP: Attention-Aware Layer-Wise Relevance Propagation for Transformers
• Faithfulness Tests for Natural Language Explanations
• Revisiting LRP: Positional Attribution as the Missing Ingredient for Transformer Explainability
• LLMs for Explainable AI: A Comprehensive Survey
• Transformer Interpretability Beyond Attention Visualization
• Reasoning Models Don't Always Say What They Think
• FaithLM: Towards Faithful Explanations for Large Language Models
• Sparse Autoencoders Find Highly Interpretable Features in Language Models
• Opportunities and Challenges in Explainable Artificial Intelligence (XAI): A Survey
• The Llama 3 Herd of Models
• DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
• Measuring Massive Multitask Language Understanding
• Can Large Language Models Explain Themselves? A Study of LLM-Generated Self-Explanations
• OpenAI o1 System Card
• Learning the Difference that Makes a Difference with Counterfactually-Augmented Data
• Are self-explanations from Large Language Models faithful?
• GPT-4 Technical Report
• GPT-4o System Card
• On Measuring Faithfulness or Self-consistency of Natural Language Explanations

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering