Interpretability papers

Browse recent papers, open the full write-up, or filter the archive by topic.

26 papers found

ReviewPaperPublishedTopics

7.2When Do LLMs Reason? A Dynamical Systems View via Entropy Phase TransitionsUnderstanding when and how to invoke reasoning in LLMs can lead to more efficient and effective AI systems, saving computational resources while improving performance.May 25, 2026LLM Reasoning Evaluation

7.5EvalMORAAL: Interpretable Chain-of-Thought and LLM-as-Judge Evaluation for Moral Alignment in Large Language ModelsUnderstanding moral alignment in AI is crucial for developing culturally aware systems that can operate effectively across diverse global contexts.May 22, 2026Alignment Evaluation Interpretability

7.2Pedestrian-Aware LLM-Driven Behavioral Planning for Autonomous VehiclesUnderstanding pedestrian behavior is crucial for the safety of autonomous vehicles in urban settings, and this approach offers a potentially more adaptable and interpretable solution than traditional methods.May 19, 2026LLM Agent Reasoning

6.8Belief Engine: Configurable and Inspectable Stance Dynamics in Multi-Agent LLM DeliberationUnderstanding and controlling stance dynamics in AI agents can enhance their reliability and transparency in applications like negotiation and conflict resolution.May 18, 2026LLM Agent Reasoning

7.2CHAL: Council of Hierarchical Agentic LanguageThis research could enhance the transparency and alignment of AI reasoning systems, making them more interpretable and subject to human oversight.May 14, 2026Agent Reasoning Alignment

6.8Agentic Interpretation: Lattice-Structured Evidence for LLM-Based Program AnalysisThis framework could improve program analysis by leveraging LLMs to incorporate dynamic information sources, potentially leading to more accurate and comprehensive analyses.May 14, 2026LLM Reasoning Coding

6.5Alignment as JurisprudenceUnderstanding the intersection of legal theory and AI alignment could lead to more robust AI systems that better conform to human values and legal standards.May 13, 2026Alignment Interpretability

7.2An Interpretable and Scalable Framework for Evaluating Large Language ModelsThis framework could enhance the evaluation of large language models by providing more scalable and interpretable insights, potentially leading to better model development and benchmark design.May 11, 2026LLM Evaluation Interpretability

7.2Understanding Annotator Safety Policy with InterpretabilityUnderstanding the reasons behind annotation disagreements can lead to more effective and inclusive AI safety policies, which is crucial for developing reliable AI systems.May 8, 2026Safety Interpretability Data

7.5Evaluating Explainability in Safety-Critical ATR Systems: Limitations of Post-Hoc Methods and Paths Toward Robust XAIUnderstanding the limitations of explainability methods in safety-critical systems is crucial for developing reliable AI applications, particularly in defense and security.May 8, 2026Safety Interpretability Evaluation

7.2LLMs Should Not Yet Be Credited with Decision ExplanationUnderstanding the limitations of LLMs in decision explanation is crucial for researchers and practitioners to avoid misinterpretations of their capabilities and to improve the development of reliable AI systems.May 6, 2026LLM Reasoning Interpretability

7.0Accurate Legal Reasoning at Scale: Neuro-Symbolic Offloading and Structural Auditability for Robust Legal AdjudicationThis approach could significantly improve the efficiency and reliability of legal reasoning systems, which are crucial for handling complex legal texts at scale.May 5, 2026LLM Reasoning Interpretability

6.8Why Do LLMs Struggle in Strategic Play? Broken Links Between Observations, Beliefs, and ActionsUnderstanding and addressing these gaps is crucial for safely deploying LLMs in strategic applications like negotiation and policymaking.May 5, 2026LLM Reasoning Interpretability

7.2Compliance versus Sensibility: On the Reasoning Controllability in Large Language ModelsUnderstanding and improving reasoning controllability in LLMs can enhance their reliability and adaptability in diverse applications, making them more useful in real-world scenarios.May 1, 2026LLM Reasoning Interpretability

7.2Evaluation of Automatic Speech Recognition Using Generative Large Language ModelsThis research provides insights into improving ASR evaluation methods, which could lead to more accurate and human-aligned speech recognition systems.Apr 30, 2026LLM Evaluation Interpretability

6.8Foundation Models in Biomedical Imaging: Turning Hype into RealityUnderstanding the practical limitations and potential of foundation models in biomedical imaging is crucial for their effective integration into clinical workflows.Apr 23, 2026Evaluation Interpretability Safety

6.8From Actions to Understanding: Conformal Interpretability of Temporal Concepts in LLM AgentsUnderstanding and improving the decision-making processes of LLM agents in interactive environments is crucial for developing trustworthy and effective autonomous systems.Apr 23, 2026LLM Agent Reasoning

7.2Reasoning Models Will Sometimes Lie About Their ReasoningUnderstanding how LRMs handle hints is crucial for improving their reliability and interpretability, especially in security-sensitive applications.Apr 22, 2026LLM Reasoning Interpretability

7.5LePREC: Reasoning as Classification over Structured Factors for Assessing Relevance of Legal IssuesThis research addresses a critical gap in legal resource accessibility by improving the precision of legal issue identification, which can aid in better legal decision-making.Apr 22, 2026LLM Reasoning Interpretability

7.5LLMbench: A Comparative Close Reading Workbench for Large Language ModelsThis tool provides researchers in the digital humanities with a new way to engage with LLM outputs, potentially enriching the understanding of generative AI's implications.Apr 21, 2026LLM Evaluation Interpretability

6.8LLM Reasoning Is Latent, Not the Chain of ThoughtUnderstanding the true nature of LLM reasoning can improve interpretability, faithfulness, and the design of reasoning benchmarks, impacting how AI systems are evaluated and developed.Apr 21, 2026LLM Reasoning Interpretability

7.5Disentangling Mathematical Reasoning in LLMs: A Methodological Investigation of Internal MechanismsUnderstanding the internal reasoning processes of LLMs can inform model design and improve their performance on reasoning-intensive tasks.Apr 20, 2026LLM Reasoning Interpretability

8.5TRUST Agents: A Collaborative Multi-Agent Framework for Fake News Detection, Explainable Verification, and Logic-Aware Claim ReasoningThis paper addresses the critical issue of fake news detection by improving the transparency and reasoning processes involved in fact verification, which is essential for building trust in automated systems.Apr 16, 2026Agent Evaluation Interpretability

8.5PRISM Risk Signal Framework: Hierarchy-Based Red Lines for AI Behavioral RiskThis paper presents a significant shift in AI safety paradigms by focusing on the underlying reasoning structures of AI systems, potentially preventing harmful outputs before they occur.Apr 14, 2026Alignment Evaluation Interpretability

8.5AI Integrity: A New Paradigm for Verifiable AI GovernanceThis paper matters because it shifts the focus from outcome-based evaluations to a procedural approach that emphasizes transparency and auditability in AI systems, potentially leading to more trustworthy AI applications in critical sectors.Apr 14, 2026Alignment Interpretability Security

8.5Principles Do Not Apply Themselves: A Hermeneutic Perspective on AI AlignmentUnderstanding the interpretive nature of applying alignment principles can lead to more robust AI systems that better reflect human values in complex situations.Apr 14, 2026Alignment Evaluation Interpretability