Results for "fixed positions"

Encodes positional information via rotation in embedding space.

Injects sequence order into Transformers, since attention alone is permutation-invariant.

Groups adopting extreme positions.

Maximum number of tokens the model can attend to in one forward pass; constrains long-document reasoning.

Search algorithm for generation that keeps top-k partial sequences; can improve likelihood but reduce diversity.

Samples from the smallest set of tokens whose probabilities sum to p, adapting set size by context.

GNN using attention to weight neighbor contributions dynamically.

Diffusion model trained to remove noise step by step.

Autoencoder using probabilistic latent variables and KL regularization.

Transformer applied to image patches.

Neural networks that operate on graph-structured data by propagating information along edges.

Internal sensing of joint positions, velocities, and forces.

Extension of convolution to graph domains using adjacency structure.

Study of motion without considering forces.

Distributed agents producing emergent intelligence.