Language Model Activations Inhabit Privileged Error-Correcting Basins
Research
Research News
New papers, breakthroughs and academic AI research
Copying Before Suppression: What Drives a Below-Chance Dip During Language Model Training?
Research
MAGEFormer: Learning Metric-Consistent Representations for Anisotropic CT Segmentation
Research
Evolving LLM-Generated Features for Interpretable Classification
Research
The Cost of a Hop: Benchmarking NLIP and A2A
Research
CUAWright: A Minimal Unified Interface for Digital Agents
Research
Agentic Cognitive Depth: Operational Criteria for Evaluating LLM Agents
Research
Evidence and Intervention: A Coupled Active-Inference Extension of Rational Speech Act Models
Research
VolS-GS: Relightable Gaussian Splatting with Volumetric Subsurface Scattering
Research
Selective Backpropagation for Efficient Few-Shot Class-Incremental Learning
Research
Agent Policy-Value Audit: Separating Transition Composition from Event Selection in Financial LLM Agents
Research
Self-Propagating Misalignment in LLM Agents, and Why Auditing or Disabling Memory Is Not Enough
Research