Chain-of-Thought reasoning in LLMs is often unfaithful, with models relying on it variably by task and less so as models scale larger.
Towards Faithfully Interpretable NLP Systems: How Should We Define and Evaluate Faithfulness?
16 Pith papers cite this work. Polarity classification is still indexing.
16
Pith papers citing it
citation-role summary
background 2
method 1
citation-polarity summary
representative citing papers
Proposes SCSuff metric for evaluating LLM explanation sufficiency via model-generated alternative inputs, showing explanations are typically insufficient and predictable from hidden states.
Pruning attention layers in five LLMs across eight datasets maintains accuracy but degrades faithfulness and calibration.
Sgatlin replaces transformer FF layers with sparse single linear neurons, improving perplexity across compute budgets and enabling direct interpretation of semantically clustered circuits for factual recall.
A new framework quantifies faithful confidence expression in large reasoning models by comparing linguistic decisiveness to token probabilities, hidden states, and response consistency, revealing it as a persistent challenge.
Summary reasoning traces from LLMs maintain task performance and increase trust and appeal relative to answer-only or full-trace conditions, but none of the formats improve users' metacognitive calibration on reasoning tasks.
Interpretability research should be judged by actionability—the degree to which its insights support concrete decisions and interventions—rather than explanatory power alone.
Counterfactual prompting effects on LLMs are often indistinguishable from those caused by meaning-preserving paraphrases, causing most previously reported demographic sensitivities to disappear under proper statistical comparison.
SLRC quantifies genuine step necessity in LLM reasoning as a causal estimator, LC-CoSR training reduces rigidity with stability guarantees, and evaluations reveal a faithfulness-sycophancy paradox across frontier models.
Joint NMF and binomial regression learns response-relevant text signals with competitive performance on simulations and review data.
ECPO is a listwise policy optimization method that couples ranking utility with span-level evidence certificate validity and a deterministic verifier reward on MAVEN-ERE and RAMS datasets.
Activation verbalization methods for LLMs largely reflect the verbalizer model's parametric knowledge rather than privileged information from the target model's activations.
LLMs support decision prediction and rationale generation but lack evidence for genuine decision explanation, requiring stricter standards to avoid over-crediting.
citing papers explorer
-
What LLMs explain is not what they believe: Evaluating explanation sufficiency under models' own input beliefs
Proposes SCSuff metric for evaluating LLM explanation sufficiency via model-generated alternative inputs, showing explanations are typically insufficient and predictable from hidden states.
-
Don't Go Breaking My LLM: The Impact of Pruning Attention Layers on Explanation Faithfulness and Confidence Calibration
Pruning attention layers in five LLMs across eight datasets maintains accuracy but degrades faithfulness and calibration.
-
Sparsely gated tiny linear experts
Sgatlin replaces transformer FF layers with sparse single linear neurons, improving perplexity across compute budgets and enabling direct interpretation of semantically clustered circuits for factual recall.
-
Quantifying Faithful Confidence Expression in Large Reasoning Models
A new framework quantifies faithful confidence expression in large reasoning models by comparing linguistic decisiveness to token probabilities, hidden states, and response consistency, revealing it as a persistent challenge.
-
Explaining Too Much? Understanding How Large Language Model Reasoning Traces Influence Performance and Metacognition
Summary reasoning traces from LLMs maintain task performance and increase trust and appeal relative to answer-only or full-trace conditions, but none of the formats improve users' metacognitive calibration on reasoning tasks.
-
Interpretability Can Be Actionable
Interpretability research should be judged by actionability—the degree to which its insights support concrete decisions and interventions—rather than explanatory power alone.
-
Compared to What? Baselines and Metrics for Counterfactual Prompting
Counterfactual prompting effects on LLMs are often indistinguishable from those caused by meaning-preserving paraphrases, causing most previously reported demographic sensitivities to disappear under proper statistical comparison.
-
Measuring and curing reasoning rigidity: from decorative chain-of-thought to genuine faithfulness
SLRC quantifies genuine step necessity in LLM reasoning as a causal estimator, LC-CoSR training reduces rigidity with stability guarantees, and evaluations reveal a faithfulness-sycophancy paradox across frontier models.
-
Learning Interpretable Text Signals for Structured Responses
Joint NMF and binomial regression learns response-relevant text signals with competitive performance on simulations and review data.
-
ECPO: Evidence-Coupled Policy Optimization for Evidence-Certified Candidate Ranking
ECPO is a listwise policy optimization method that couples ranking utility with span-level evidence certificate validity and a deterministic verifier reward on MAVEN-ERE and RAMS datasets.
-
LLMs Should Not Yet Be Credited with Decision Explanation
LLMs support decision prediction and rationale generation but lack evidence for genuine decision explanation, requiring stricter standards to avoid over-crediting.
- NEURON: A Neuro-symbolic System for Grounded Clinical Explainability
- Evaluating LLM-Driven Summarisation of Parliamentary Debates with Computational Argumentation
- Beyond Explainable AI (XAI): An Overdue Paradigm Shift and Post-XAI Research Directions