#AI Reliability

Articles (17)

• 🇺🇸 Breaking the Chain: A Causal Analysis of LLM Faithfulness to Intermediate Structures — 18/03/2026 [USA]

  arXiv:2603.16475v1 Announce Type: new Abstract: Schema-guided reasoning pipelines ask LLMs to produce explicit intermediate structures -- rubrics, checklists, verification queries -- before committin...

  Related: #Causal Analysis
• 🇺🇸 Grounding the Score: Explicit Visual Premise Verification for Reliable Vision-Language Process Reward Models — 18/03/2026 [USA]

  arXiv:2603.16253v1 Announce Type: cross Abstract: Vision-language process reward models (VL-PRMs) are increasingly used to score intermediate reasoning steps and rerank candidates under test-time sca...

  Related: #Multimodal Verification
• 🇺🇸 When Stability Fails: Hidden Failure Modes Of LLMS in Data-Constrained Scientific Decision-Making — 18/03/2026 [USA]

  arXiv:2603.15840v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used as decision-support tools in data-constrained scientific workflows, where correctness and validity...

  Related: #Scientific Research
• 🇺🇸 Attribution-Guided Model Rectification of Unreliable Neural Network Behaviors — 18/03/2026 [USA]

  arXiv:2603.15656v1 Announce Type: cross Abstract: The performance of neural network models deteriorates due to their unreliable behavior on non-robust features of corrupted samples. Owing to their op...

  Related: #Model Correction
• 🇺🇸 Nonstandard Errors in AI Agents — 18/03/2026 [USA]

  arXiv:2603.16744v1 Announce Type: new Abstract: We study whether state-of-the-art AI coding agents, given the same data and research question, produce the same empirical results. Deploying 150 autono...

  Related: #Error Analysis
• 🇺🇸 Thinking in Uncertainty: Mitigating Hallucinations in MLRMs with Latent Entropy-Aware Decoding — 17/03/2026 [USA]

  arXiv:2603.13366v1 Announce Type: cross Abstract: Recent advancements in multimodal large reasoning models (MLRMs) have significantly improved performance in visual question answering. However, we ob...

  Related: #Uncertainty Management
• 🇺🇸 Semantic Invariance in Agentic AI — 16/03/2026 [USA]

  arXiv:2603.13173v1 Announce Type: new Abstract: Large Language Models (LLMs) increasingly serve as autonomous reasoning agents in decision support, scientific problem-solving, and multi-agent coordin...

  Related: #Natural Language Processing
• 🇺🇸 Deployment-Time Reliability of Learned Robot Policies — 13/03/2026 [USA]

  arXiv:2603.11400v1 Announce Type: cross Abstract: Recent advances in learning-based robot manipulation have produced policies with remarkable capabilities. Yet, reliability at deployment remains a fu...

  Related: #Robotics
• 🇺🇸 Toward Epistemic Stability: Engineering Consistent Procedures for Industrial LLM Hallucination Reduction — 12/03/2026 [USA]

  arXiv:2603.10047v1 Announce Type: cross Abstract: Hallucinations in large language models (LLMs) are outputs that are syntactically coherent but factually incorrect or contextually inconsistent. They...

  Related: #Industrial AI
• 🇺🇸 Quantifying Hallucinations in Language Language Models on Medical Textbooks — 12/03/2026 [USA]

  arXiv:2603.09986v1 Announce Type: cross Abstract: Hallucinations, the tendency for large language models to provide responses with factually incorrect and unsupported claims, is a serious problem wit...

  Related: #Medical AI
• 🇺🇸 Consistency-based Abductive Reasoning over Perceptual Errors of Multiple Pre-trained Models in Novel Environments — 12/03/2026 [USA]

  arXiv:2505.19361v5 Announce Type: replace Abstract: The deployment of pre-trained perception models in novel environments often leads to performance degradation due to distributional shifts. Although...

  Related: #Error Detection
• 🇺🇸 Amazon plans 'deep dive' internal meeting to address AI-related outages — 10/03/2026 [USA]

  Amazon said AI-assisted production changes were partly to blame for recent infrastructure issues.

  Related: #Corporate Strategy
• 🇺🇸 Make VLM Recognize Visual Hallucination on Cartoon Character Image with Pose Information — 09/03/2026 [USA]

  arXiv:2403.15048v4 Announce Type: replace-cross Abstract: Leveraging large-scale Text-to-Image (TTI) models have become a common technique for generating exemplar or training dataset in the fields of...

  Related: #Computer Vision
• 🇺🇸 On the Reliability of AI Methods in Drug Discovery: Evaluation of Boltz-2 for Structure and Binding Affinity Prediction — 09/03/2026 [USA]

  arXiv:2603.05532v1 Announce Type: cross Abstract: Despite continuing hype about the role of AI in drug discovery, no "AI-discovered drugs" have so far received regulatory approval. Here we assess one...

  Related: #Drug Discovery, #Protein Prediction
• 🇺🇸 Induced Numerical Instability: Hidden Costs in Multimodal Large Language Models — 06/03/2026 [USA]

  arXiv:2603.04453v1 Announce Type: cross Abstract: The use of multimodal large language models has become widespread, and as such the study of these models and their failure points has become of utmos...

  Related: #Computational Costs
• 🇺🇸 Agent Behavioral Contracts: Formal Specification and Runtime Enforcement for Reliable Autonomous AI Agents — 27/02/2026 [USA]

  arXiv:2602.22302v1 Announce Type: new Abstract: Traditional software relies on contracts -- APIs, type systems, assertions -- to specify and enforce correct behavior. AI agents, by contrast, operate ...

  Related: #Formal Specification, #Runtime Enforcement
• 🇺🇸 Oracular Programming: A Modular Foundation for Building LLM-Enabled Software — 25/02/2026 [USA]

  arXiv:2502.05310v4 Announce Type: replace-cross Abstract: Large Language Models can solve a wide range of tasks from just a few examples, but they remain difficult to steer and lack a capability esse...

  Related: #Programming Paradigms, #LLM Integration, #Software Architecture