internal representations
The encoded features or attributes that a model learns from the data during training, which capture underlying patterns and facilitate decision-making or output generation.
- Bridging Critical Gaps in Convergent Learning: How Representational Alignment Evolves Across Layers, Training, and Distribution Shifts
- Continuous Thought Machines
- Cross-modal Associations in Vision and Language Models: Revisiting the Bouba-Kiki Effect
- Learning Diffusion Models with Flexible Representation Guidance
- On Optimal Steering to Achieve Exact Fairness
- Optimization Inspired Few-Shot Adaptation for Large Language Models
- Robust Hallucination Detection in LLMs via Adaptive Token Selection
- When Semantics Mislead Vision: Mitigating Large Multimodal Models Hallucinations in Scene Text Spotting and Understanding