Explainable Artificial Intelligence (XAI)

Derniers papiers

• DeepLeak: Privacy Enhancing Hardening of Model Explanations Against Membership Leakage

  Firas Ben Hmida, Zain Sbeih, Philemon Hailemariam, Birhanu Eshete · 8 janvier 2026

  Machine learning (ML) explainability is central to algorithmic transparency in high-stakes settings such as predictive diagnostics and loan approval. However, these same domains require rigorous privacy guaranties, creating tension between interpretability and privacy. Although prior work has shown …

• Compact Example-Based Explanations for Language Models

  Loris Schoenegger, Benjamin Roth · 8 janvier 2026

  Training data influence estimation methods quantify the contribution of training documents to a model's output, making them a promising source of information for example-based explanations. As humans cannot interpret thousands of documents, only a small subset of the training data can be presented a…

• In Search of Grandmother Cells: Tracing Interpretable Neurons in Tabular Representations

  Ricardo Knauer, Erik Rodner · 8 janvier 2026

  Foundation models are powerful yet often opaque in their decision-making. A topic of continued interest in both neuroscience and artificial intelligence is whether some neurons behave like grandmother cells, i.e., neurons that are inherently interpretable because they exclusively respond to single c…

• Shielded RecRL: Explanation Generation for Recommender Systems without Ranking Degradation

  Ansh Tiwari, Ayush Chauhan · 8 janvier 2026

  We introduce Shielded RecRL, a reinforcement learning approach to generate personalized explanations for recommender systems without sacrificing the system's original ranking performance. Unlike prior RLHF-based recommender methods that directly optimize item rankings, our two-tower architecture kee…

• A Gap Between Decision Trees and Neural Networks

  Akash Kumar · 8 janvier 2026

  We study when geometric simplicity of decision boundaries, used here as a notion of interpretability, can conflict with accurate approximation of axis-aligned decision trees by shallow neural networks. Decision trees induce rule-based, axis-aligned decision regions (finite unions of boxes), whereas …

• The Cost of Dynamic Reasoning: Demystifying AI Agents and Test-Time Scaling from an AI Infrastructure Perspective

  Jiin Kim, Byeongjun Shin, Jinha Chung, Minsoo Rhu · 8 janvier 2026

  Large-language-model (LLM)-based AI agents have recently showcased impressive versatility by employing dynamic reasoning, an adaptive, multi-step process that coordinates with external tools. This shift from static, single-turn inference to agentic, multi-turn workflows broadens task generalization …

• Improving Compactness and Reducing Ambiguity of CFIRE Rule-Based Explanations

  Sebastian M\"uller, Tobias Schneider, Ruben Kemna, Vanessa Toborek · 8 janvier 2026

  Models trained on tabular data are widely used in sensitive domains, increasing the demand for explanation methods to meet transparency needs. CFIRE is a recent algorithm in this domain that constructs compact surrogate rule models from local explanations. While effective, CFIRE may assign rules ass…

• Causal Data Augmentation for Robust Fine-Tuning of Tabular Foundation Models

  Magnus B\"uhler, Lennart Purucker, Frank Hutter · 8 janvier 2026

  Fine-tuning tabular foundation models (TFMs) under data scarcity is challenging, as early stopping on even scarcer validation data often fails to capture true generalization performance. We propose CausalMixFT, a method that enhances fine-tuning robustness and downstream performance by generating st…

• NeuronScope: A Multi-Agent Framework for Explaining Polysemantic Neurons in Language Models

  Weiqi Liu, Yongliang Miao, Haiyan Zhao, Yanguang Liu, Mengnan Du · 8 janvier 2026

  Neuron-level interpretation in large language models (LLMs) is fundamentally challenged by widespread polysemanticity, where individual neurons respond to multiple distinct semantic concepts. Existing single-pass interpretation methods struggle to faithfully capture such multi-concept behavior. In t…

• Mechanistic Knobs in LLMs: Retrieving and Steering High-Order Semantic Features via Sparse Autoencoders

  Ruikang Zhang, Shuo Wang, Qi Su · 7 janvier 2026

  Recent work in Mechanistic Interpretability (MI) has enabled the identification and intervention of internal features in Large Language Models (LLMs). However, a persistent challenge lies in linking such internal features to the reliable control of complex, behavior-level semantic attributes in lang…

• Lil: Less is Less When Applying Post-Training Sparse-Attention Algorithms in Long-Decode Stage

  Junhao Hu, Fangze Li, Mingtao Xu, Feifan Meng, Shiju Zhao, Tiancheng Hu, Ting Peng, Anmin Liu, Wenrui Huang, Chenxu Liu, Ziyue Hua, Tao Xie · 7 janvier 2026

  Large language models (LLMs) demonstrate strong capabilities across a wide range of complex tasks and are increasingly deployed at scale, placing significant demands on inference efficiency. Prior work typically decomposes inference into prefill and decode stages, with the decode stage dominating to…

• The Vibe-Check Protocol: Quantifying Cognitive Offloading in AI Programming

  Aizierjiang Aiersilan · 7 janvier 2026

  The integration of Large Language Models (LLMs) into software engineering education has driven the emergence of ``Vibe Coding,'' a paradigm where developers articulate high-level intent through natural language and delegate implementation to AI agents. While proponents argue this approach modernizes…

• InfiAgent: An Infinite-Horizon Framework for General-Purpose Autonomous Agents

  Chenglin Yu, Yuchen Wang, Songmiao Wang, Hongxia Yang, Ming Li · 7 janvier 2026

  LLM agents can reason and use tools, but they often break down on long-horizon tasks due to unbounded context growth and accumulated errors. Common remedies such as context compression or retrieval-augmented prompting introduce trade-offs between information fidelity and reasoning stability. We pres…

• The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systems

  Richard Ren, Arunim Agarwal, Mantas Mazeika, Cristina Menghini, Robert Vacareanu, Brad Kenstler, Mick Yang, Isabelle Barrass, Alice Gatti, Xuwang Yin, Eduardo Trevino, Matias Geralnik, Adam Khoja, Dean Lee, Summer Yue, Dan Hendrycks · 7 janvier 2026

  As large language models (LLMs) become more capable and agentic, the requirement for trust in their outputs grows significantly, yet at the same time concerns have been mounting that models may learn to lie in pursuit of their goals. To address these concerns, a body of work has emerged around the n…

• Towards Trustworthy LLM-Based Recommendation via Rationale Integration

  Chung Park, Taesan Kim, Hyeongjun Yun, Dongjoon Hong, Junui Hong, Kijung Park, MinCheol Cho, Mira Myong, Jihoon Oh, Min sung Choi · 7 janvier 2026

  Traditional recommender systems (RS) have been primarily optimized for accuracy and short-term engagement, often overlooking transparency and trustworthiness. Recently, platforms such as Amazon and Instagram have begun providing recommendation rationales to users, acknowledging their critical role i…

• Textual Explanations and Their Evaluations for Reinforcement Learning Policy

  Ahmad Terra, Mohit Ahmed, Rafia Inam, Elena Fersman, Martin T\"orngren · 7 janvier 2026

  Understanding a Reinforcement Learning (RL) policy is crucial for ensuring that autonomous agents behave according to human expectations. This goal can be achieved using Explainable Reinforcement Learning (XRL) techniques. Although textual explanations are easily understood by humans, ensuring their…

• Reconsidering Overthinking: Penalizing Internal and External Redundancy in CoT Reasoning

  Jialiang Hong, Taihang Zhen, Kai Chen, Jiaheng Liu, Junlan Feng, Wenpeng Zhu, Jing Huo, Yang Gao, Depeng Wang, Haitao Wan, Xi Yang, Boyan Wang, Fanyu Meng, Yuyao Zhang · 7 janvier 2026

  Large Reasoning Models (LRMs) often suffer from overthinking, generating verbose reasoning traces that compromise both computational efficiency and interpretability. Unlike prior efforts that rely on global length-based rewards, we propose a semantic-aware decomposition of redundancy into two distin…

• Prompt-Counterfactual Explanations for Generative AI System Behavior

  Sofie Goethals, Foster Provost, Jo\~ao Sedoc · 7 janvier 2026

  As generative AI systems become integrated into real-world applications, organizations increasingly need to be able to understand and interpret their behavior. In particular, decision-makers need to understand what causes generative AI systems to exhibit specific output characteristics. Within this …

• When the Coffee Feature Activates on Coffins: An Analysis of Feature Extraction and Steering for Mechanistic Interpretability

  Raphael Ronge, Markus Maier, Frederick Eberhardt · 7 janvier 2026

  Recent work by Anthropic on Mechanistic interpretability claims to understand and control Large Language Models by extracting human-interpretable features from their neural activation patterns using sparse autoencoders (SAEs). If successful, this approach offers one of the most promising routes for …

• iFlip: Iterative Feedback-driven Counterfactual Example Refinement

  Yilong Wang, Qianli Wang, Nils Feldhus · 6 janvier 2026

  Counterfactual examples are minimal edits to an input that alter a model's prediction. They are widely employed in explainable AI to probe model behavior and in natural language processing (NLP) to augment training data. However, generating valid counterfactuals with large language models (LLMs) rem…

• Horizon Activation Mapping for Neural Networks in Time Series Forecasting

  Hans Krupakar, V A Kandappan · 6 janvier 2026

  Neural networks for time series forecasting have relied on error metrics and architecture-specific interpretability approaches for model selection that don't apply across models of different families. To interpret forecasting models agnostic to the types of layers across state-of-the-art model famil…

• Aletheia: Quantifying Cognitive Conviction in Reasoning Models via Regularized Inverse Confusion Matrix

  Fanzhe Fu · 6 janvier 2026

  In the progressive journey toward Artificial General Intelligence (AGI), current evaluation paradigms face an epistemological crisis. Static benchmarks measure knowledge breadth but fail to quantify the depth of belief. While Simhi et al. (2025) defined the CHOKE phenomenon in standard QA, we extend…

• Empowering Small Language Models with Factual Hallucination-Aware Reasoning for Financial Classification

  Han Yuan, Yilin Wu, Li Zhang, Zheng Ma · 6 janvier 2026

  Small language models (SLMs) are increasingly used for financial classification due to their fast inference and local deployability. However, compared with large language models, SLMs are more prone to factual hallucinations in reasoning and exhibit weaker classification performance. This raises a n…

• A three-Level Framework for LLM-Enhanced eXplainable AI: From technical explanations to natural language

  Marilyn Bello, Rafael Bello, Maria-Matilde Garc\'ia, Ann Now\'e, Iv\'an Sevillano-Garc\'ia, Francisco Herrera · 6 janvier 2026

  The growing application of artificial intelligence in sensitive domains has intensified the demand for systems that are not only accurate but also explainable and trustworthy. Although explainable AI (XAI) methods have proliferated, many do not consider the diverse audiences that interact with AI sy…

• Admissibility Alignment

  Chris Duffey · 6 janvier 2026

  This paper introduces Admissibility Alignment: a reframing of AI alignment as a property of admissible action and decision selection over distributions of outcomes under uncertainty, evaluated through the behavior of candidate policies. We present MAP-AI (Monte Carlo Alignment for Policy) as a canon…