One-Step Curvature Probes Miss the Fitting Operator: Retained Capacity and Terminal Null-Space Correction for Continual Learning
Research
Research News
New papers, breakthroughs and academic AI research
The Independence Prior of SAEs Fragments Visual Concepts
Research
Learning Latent Protein Languages for Autoregressive Generation
Research
Reinforcement Learning with Comparative Evidence for Social Intelligence
Research
Agentic AI with Structured CoT for Enhancing AI's Spatial Intelligence: Visualization and Reasoning of Rotation
Research
Towards Safer Autonomous Driving in an Open World: A Dual-Process Approach
Research
Auditing Pairwise Equivalence Judgments: Self-Critique Effects and Diversity Measurement in Multi-Agent Hypothesis Generation
Research
SHarP: Saliency-based Pruning of Agent Harnesses
Research
EvalResearchBench: Can AI Agents Design Their Own Evaluations?
Research
Playing social deduction games with reinforcement fine-tuned large language models
Research
Dynamic Quadtree Tokenization and Transformer for Adaptive Mesh PDE Forecasting
Research
AI-Enabled Quality Assurance for Multiple-Choice Assessment Items
Research