representational alignment
Representational alignment refers to the degree to which different representations (e.g., learned embeddings) correspond with one another across different domains or tasks, facilitating transfer learning and collaboration between models.
- A Closer Look at Graph Transformers: Cross-Aggregation and Beyond
- Beyond Scalars: Concept-Based Alignment Analysis in Vision Transformers
- Meta-Learning an In-Context Transformer Model of Human Higher Visual Cortex
- Model–Behavior Alignment under Flexible Evaluation: When the Best-Fitting Model Isn’t the Right One
- Quantifying Task-relevant Similarities in Representations Using Decision Variable Correlations
- Scaling and context steer LLMs along the same computational path as the human brain