On the estimation and validity of AI time horizons—a statistical look at the METR plot
UC Berkeley
BrickBench: Evaluating Agentic Brick Design
Stanford University; Max Planck Institute for Intelligent Systems; Inria
Ecology of AI Agents: Collaboration Creates a Population Threshold for Takeoff
Harvard University; NTT Research, Inc.
Searching for “Harmful Refusal”: A Psychometric Audit of an AI Safety Benchmark
Carnegie Mellon University; Indiana University; Google
HRIL: Learning Multimodal Synergy via Higher-Order Tensor Modeling
Southwestern University of Finance and Economics; Artificial Intelligence and Digital Finance Key Laboratory of Sichuan Province; X-Humanoid; City University of Hong Kong; Zhejiang Sci-Tech University; Chinese Academy of Sciences; Chengdu Everimaging Science and Technology Co., Ltd; University of Electronic Science and Technology of China
GeoReform: Reflective Formalization Evolution for Multimodal Geometry Problem Solving
Tongji University; University of North Carolina at Chapel Hill; Arizona State University
OnTrack: Real-Time Monitoring and Intervention in LLM Agent Trajectories via Streaming Structure-Aware Optimal Transport
Cribl AI Research Lab
HANS: A Handwritten Answer Sheet Dataset for Noisy Hybrid Document Parsing
Jinan University
Cited but Not Consulted: A Counterfactual Audit of Legal Chain-of-Thought Faithfulness
Lexsi Labs
Accurate but Not Humble: Evaluating Epistemic Humility in LLM Agents under Knowledge Conflict
Johns Hopkins University; New York University
Overcoming Prior Barriers: Supervised Fine-Tuning under Long-Tail Distribution
Virginia Tech; Amazon; Meta; MBZUAI; Dartmouth College
Can AI Agents Learn Their Way to the Top? Evaluating Heuristic Learning in a Long-Running Game Agent Competition
Tsinghua University
Prior or Feedback? What an LLM Uses When Adapting Neural Operators
University of Surrey; Compare the Market
Verdict Without the Rule: Diagnosing and Auditing Regulatory Rule Sensitivity in LLM Compliance Systems
Lexsi Labs
A Structural Theory of Cognitive Representation and Problem Solving,Contexts, Invariance, and the Knowledge Space
HUN-REN Wigner Research Centre for Physics; Corvinus University of Budapest
Looking Inside LLMs: Small-World Connectivity as a Signature of Reasoning Performance
Dartmouth College; Yale University; New York University; UNC Charlotte; Virginia Tech
Learning Probabilistic Logic Programs with Functional Gradient Guided Language Models
Technische Universität Darmstadt; The University of Texas at Dallas; AT&T CDO; Oregon state University
One Word Opens the Gate: The Option-Channel Attack on Typed Decision Models as Agent Guardrails
University of Southern California (USC)
When Has a Bayesian Neural Network Sampled Enough? Adaptive Inference Time with Statistical Guarantees
Eindhoven University of Technology
Recursive Self-Improvement through Multi-Agent Self-Supervision
UC Berkeley; Sakana AI