06:00How RLHF amplifies sycophancy in language modelsResearch & Sciencecs.AI updates on arXiv.org9/10Relevance 9 out of 10
06:00ECPO: Evidence-Coupled Policy Optimization for Verifier-Audited Candidate RankingResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00ATWL: A formal language for modeling visual analytics workflowsResearch & Sciencecs.AI updates on arXiv.org7/10Relevance 7 out of 10
06:00Compositional prompt optimization with discrete codebooksResearch & Sciencecs.AI updates on arXiv.org9/10Relevance 9 out of 10
06:00Multimodal models fail to detect implicit harmful content from image-text compositionResearch & Sciencecs.AI updates on arXiv.org9/10Relevance 9 out of 10
06:00Large language models in transportation systems management and operationsResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00Moral sycophancy in vision-language modelsResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00Sparse Autoencoders to Mitigate Catastrophic Forgetting in Continual LearningResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00DriveWorld3D: 3D World Modeling with Evidence Constraints for Autonomous DrivingResearch & Sciencecs.AI updates on arXiv.org9/10Relevance 9 out of 10
06:00Analysis of political narratives during Moldova's 2024 presidential electionResearch & Sciencecs.AI updates on arXiv.org7/10Relevance 7 out of 10
06:00Recursive self-distillation risks progressive misalignment in AI modelsResearch & Sciencecs.AI updates on arXiv.org9/10Relevance 9 out of 10
06:00Upper bounds on the number of causal pieces in spiking neural networksResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00Mitigating demographic disparities in explanation fairness for hate speech moderationResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00Can LLMs predict the outcomes of machine learning experiments?Research & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00Tree-species recognition with LiDAR: limited cross-source generalization revealedResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00StratFL: New defense against model poisoning attacks in federated learningResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00Freeway speed forecasting with sparse sensors using graph networksResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00Plausible explanations mislead multimodal AI judgesResearch & Sciencecs.AI updates on arXiv.org9/10Relevance 9 out of 10
06:00ToMStories: Evaluating Theory of Mind in Language ModelsResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00New method to measure LLM uncertainty through causal robustnessResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00On-Demand Hard Instances for Algorithm EvaluationResearch & Sciencecs.AI updates on arXiv.org9/10Relevance 9 out of 10
06:00Diverse Reasoning Paths Matter: Symbolic Equivalence Augmentation for RL ExplorationResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00NoveltyScope: Multi-agent research assistant for novelty assessment of scientific papersResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00ARBOR: Detecting reasoning failures in long chains of thoughtResearch & Sciencecs.AI updates on arXiv.org9/10Relevance 9 out of 10
06:00OrderFlowDiT: Diffusion Transformer for Limit-Order Volume ForecastingResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00CADENCE: Adaptive decoding method to speed up text generationResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00Reducing private data leakage in fine-tuned LLMs via neuron pruningResearch & Sciencecs.AI updates on arXiv.org9/10Relevance 9 out of 10
06:00STRIDE: New benchmark for spatiotemporal reasoning in autonomous drivingResearch & Sciencecs.AI updates on arXiv.org9/10Relevance 9 out of 10
06:00The Reliance Hypothesis: Chain-of-Thought Faithfulness Rises as Language Resource Level FallsResearch & Sciencecs.AI updates on arXiv.org8/10Relevance 8 out of 10
06:00AnchorQ: Efficient Policy Improvement for Robotics via Structured Latent Action AnchorsResearch & Sciencecs.AI updates on arXiv.org9/10Relevance 9 out of 10