Explainable Artificial Intelligence (XAI)

Derniers papiers

• Quantifying the Necessity of Chain of Thought through Opaque Serial Depth

  Jonah Brown-Cohen, David Lindner, Rohin Shah · 11 mars 2026

  Large language models (LLMs) tend to externalize their reasoning in their chain of thought, making the chain of thought a good target for monitoring. This is partially an inherent feature of the Transformer architecture: sufficiently long serial cognition must pass through the chain of thought (Korb…

• From Data Statistics to Feature Geometry: How Correlations Shape Superposition

  Lucas Prieto, Edward Stevinson, Melih Barsbey, Tolga Birdal, Pedro A. M. Mediano · 11 mars 2026

  A central idea in mechanistic interpretability is that neural networks represent more features than they have dimensions, arranging them in superposition to form an over-complete basis. This framing has been influential, motivating dictionary learning approaches such as sparse autoencoders. However,…

• What is Missing? Explaining Neurons Activated by Absent Concepts

  Robin Hesse, Simone Schaub-Meyer, Janina Hesse, Bernt Schiele, Stefan Roth · 11 mars 2026

  Explainable artificial intelligence (XAI) aims to provide human-interpretable insights into the behavior of deep neural networks (DNNs), typically by estimating a simplified causal structure of the model. In existing work, this causal structure often includes relationships where the presence of a co…

• Hindsight Credit Assignment for Long-Horizon LLM Agents

  Hui-Ze Tan, Xiao-Wen Yang, Hao Chen, Jie-Jing Shao, Yi Wen, Yuteng Shen, Weihong Luo, Xiku Du, Lan-Zhe Guo, Yu-Feng Li · 11 mars 2026

  Large Language Model (LLM) agents often face significant credit assignment challenges in long-horizon, multi-step tasks due to sparse rewards. Existing value-free methods, such as Group Relative Policy Optimization (GRPO), encounter two fundamental bottlenecks: inaccurate step-level Q-value estimati…

• Beyond Scaling: Assessing Strategic Reasoning and Rapid Decision-Making Capability of LLMs in Zero-sum Environments

  Yang Li, Xing Chen, Yutao Liu, Gege Qi, Yanxian BI, Zizhe Wang, Yunjian Zhang, Yao Zhu · 11 mars 2026

  Large Language Models (LLMs) have achieved strong performance on static reasoning benchmarks, yet their effectiveness as interactive agents operating in adversarial, time-sensitive environments remains poorly understood. Existing evaluations largely treat reasoning as a single-shot capability, overl…

• The Reasoning Trap -- Logical Reasoning as a Mechanistic Pathway to Situational Awareness

  Subramanyam Sahoo, Aman Chadha, Vinija Jain, Divya Chaudhary · 11 mars 2026

  Situational awareness, the capacity of an AI system to recognize its own nature, understand its training and deployment context, and reason strategically about its circumstances, is widely considered among the most dangerous emergent capabilities in advanced AI systems. Separately, a growing researc…

• HyConEx: Hypernetwork classifier with counterfactual explanations for tabular data

  Patryk Marsza{\l}ek, Kamil Ksi\k{a}\.zek, Oleksii Furman, Ulvi Movsum-zada, Przemys{\l}aw Spurek, Marek \'Smieja · 11 mars 2026

  In recent years, there has been a growing interest in explainable AI methods. In addition to making accurate predictions, we also want to understand what the model's decision is based on. One of the fundamental levels of interpretability is to provide counterfactual examples explaining the rationale…

• CREDO: Epistemic-Aware Conformalized Credal Envelopes for Regression

  Luben M. C. Cabezas, Sabina J. Sloman, Bruno M. Resende, Fanyi Wu, Michele Caprio, Rafael Izbicki · 10 mars 2026

  Conformal prediction delivers prediction intervals with distribution-free coverage, but its intervals can look overconfident in regions where the model is extrapolating, because standard conformal scores do not explicitly represent epistemic uncertainty. Credal methods, by contrast, make epistemic e…

• Enhancing SHAP Explainability for Diagnostic and Prognostic ML Models in Alzheimer Disease

  Pablo Guill\'en, Enrique Frias-Martinez · 10 mars 2026

  Alzheimer disease (AD) diagnosis and prognosis increasingly rely on machine learning (ML) models. Although these models provide good results, clinical adoption is limited by the need for technical expertise and the lack of trustworthy and consistent model explanations. SHAP (SHapley Additive exPlana…

• Interpretable-by-Design Transformers via Architectural Stream Independence

  Clayton Kerce, Alexis Fox · 10 mars 2026

  While transformers achieve strong performance, their internal decision-making processes remain opaque. We investigate whether architectural constraints can enforce interpretability by design through architectural stream independence: maintaining a token stream (carrying symbolic structure) and conte…

• The Dual-Stream Transformer: Channelized Architecture for Interpretable Language Modeling

  J. Clayton Kerce, Alexis Fox · 10 mars 2026

  Standard transformers entangle all computation in a single residual stream, obscuring which components perform which functions. We introduce the Dual-Stream Transformer, which decomposes the residual stream into two functionally distinct components: a token stream updated by attention and a context …

• More Bang for the Buck: Process Reward Modeling with Entropy-Driven Uncertainty

  Lang Cao, Renhong Chen, Yingtian Zou, Chao Peng, Huacong Xu, Yuxian Wang, Wu Ning, Qian Chen, Mofan Peng, Zijie Chen, Peishuo Su, Yitong Li · 10 mars 2026

  We introduce the Entropy-Driven Uncertainty Process Reward Model (EDU-PRM), a novel entropy-driven training framework for process reward modeling that enables dynamic, uncertainty-aligned segmentation of complex reasoning steps, eliminating the need for costly manual step annotations. Unlike previou…

• Consensus is Not Verification: Why Crowd Wisdom Strategies Fail for LLM Truthfulness

  Yegor Denisov-Blanch, Joshua Kazdan, Jessica Chudnovsky, Rylan Schaeffer, Sheng Guan, Soji Adeshina, Sanmi Koyejo · 10 mars 2026

  Pass@k and other methods of scaling inference compute can improve language model performance in domains with external verifiers, including mathematics and code, where incorrect candidates can be filtered reliably. This raises a natural question: can we similarly scale compute to elicit gains in trut…

• DC-W2S: Dual-Consensus Weak-to-Strong Training for Reliable Process Reward Modeling in Biological Reasoning

  Chi-Min Chan, Ehsan Hajiramezanali, Xiner Li, Edward De Brouwer, Carl Edwards, Wei Xue, Sirui Han, Yike Guo, Gabriele Scalia · 10 mars 2026

  In scientific reasoning tasks, the veracity of the reasoning process is as critical as the final outcome. While Process Reward Models (PRMs) offer a solution to the coarse-grained supervision problems inherent in Outcome Reward Models (ORMs), their deployment is hindered by the prohibitive cost of o…

• TT-Sparse: Learning Sparse Rule Models with Differentiable Truth Tables

  Hans Farrell Soegeng, Sarthak Ketanbhai Modi, Thomas Peyrin · 10 mars 2026

  Interpretable machine learning is essential in high-stakes domains where decision-making requires accountability, transparency, and trust. While rule-based models offer global and exact interpretability, learning rule sets that simultaneously achieve high predictive performance and low, human-unders…

• Hierarchical Latent Structures in Data Generation Process Unify Mechanistic Phenomena across Scale

  Jonas Rohweder, Subhabrata Dutta, Iryna Gurevych · 10 mars 2026

  Contemporary studies have uncovered many puzzling phenomena in the neural information processing of Transformer-based language models. Building a robust, unified understanding of these phenomena requires disassembling a model within the scope of its training. While the intractable scale of pretraini…

• Learning Concept Bottleneck Models from Mechanistic Explanations

  Antonio De Santis, Schrasing Tong, Marco Brambilla, Lalana Kagal · 10 mars 2026

  Concept Bottleneck Models (CBMs) aim for ante-hoc interpretability by learning a bottleneck layer that predicts interpretable concepts before the decision. State-of-the-art approaches typically select which concepts to learn via human specification, open knowledge graphs, prompting an LLM, or using …

• Mitigating Bias in Concept Bottleneck Models for Fair and Interpretable Image Classification

  Schrasing Tong, Antoine Salaun, Vincent Yuan, Annabel Adeyeri, Lalana Kagal · 9 mars 2026

  Ensuring fairness in image classification prevents models from perpetuating and amplifying bias. Concept bottleneck models (CBMs) map images to high-level, human-interpretable concepts before making predictions via a sparse, one-layer classifier. This structure enhances interpretability and, in theo…

• RM-R1: Reward Modeling as Reasoning

  Xiusi Chen, Gaotang Li, Ziqi Wang, Bowen Jin, Cheng Qian, Yu Wang, Hongru Wang, Yu Zhang, Denghui Zhang, Tong Zhang, Hanghang Tong, Heng Ji · 9 mars 2026

  Reward modeling is essential for aligning large language models with human preferences through reinforcement learning. To provide accurate reward signals, a reward model (RM) should stimulate deep thinking and conduct interpretable reasoning before assigning a score or a judgment. Inspired by recent…

• From Entropy to Calibrated Uncertainty: Training Language Models to Reason About Uncertainty

  Azza Jenane, Nassim Walha, Lukas Kuhn, Florian Buettner · 9 mars 2026

  Large Language Models (LLMs) that can express interpretable and calibrated uncertainty are crucial in high-stakes domains. While methods to compute uncertainty post-hoc exist, they are often sampling-based and therefore computationally expensive or lack calibration. We propose a three-stage pipeline…

• Predictive Coding Graphs are a Superset of Feedforward Neural Networks

  Bj\"orn van Zwol · 9 mars 2026

  Predictive coding graphs (PCGs) are a recently introduced generalization to predictive coding networks, a neuroscience-inspired probabilistic latent variable model. Here, we prove how PCGs define a mathematical superset of feedforward artificial neural networks (multilayer perceptrons). This positio…

• CLAIRE: Compressed Latent Autoencoder for Industrial Representation and Evaluation -- A Deep Learning Framework for Smart Manufacturing

  Mohammadhossein Ghahramani, Mengchu Zhou · 9 mars 2026

  Accurate fault detection in high-dimensional industrial environments remains a major challenge due to the inherent complexity, noise, and redundancy in sensor data. This paper introduces CLAIRE, i.e., a hybrid end-to-end learning framework that integrates unsupervised deep representation learning wi…

• Evaluating LLM Alignment With Human Trust Models

  Anushka Debnath, Stephen Cranefield, Bastin Tony Roy Savarimuthu, Emiliano Lorini · 9 mars 2026

  Trust plays a pivotal role in enabling effective cooperation, reducing uncertainty, and guiding decision-making in both human interactions and multi-agent systems. Although it is significant, there is limited understanding of how large language models (LLMs) internally conceptualize and reason about…

• PONTE: Personalized Orchestration for Natural Language Trustworthy Explanations

  Vittoria Vineis, Matteo Silvestri, Lorenzo Antonelli, Filippo Betello, Gabriele Tolomei · 9 mars 2026

  Explainable Artificial Intelligence (XAI) seeks to enhance the transparency and accountability of machine learning systems, yet most methods follow a one-size-fits-all paradigm that neglects user differences in expertise, goals, and cognitive needs. Although Large Language Models can translate techn…

• XAI for Coding Agent Failures: Transforming Raw Execution Traces into Actionable Insights

  Arun Joshi · 9 mars 2026

  Large Language Model (LLM)-based coding agents show promise in automating software development tasks, yet they frequently fail in ways that are difficult for developers to understand and debug. While general-purpose LLMs like GPT can provide ad-hoc explanations of failures, raw execution traces rema…