Kanishk Awadhiya

H-Res achieves a 26% improvement in associative retrieval tasks by steering token trajectories without altering the model's global equilibrium.

Correct reasoning in LLMs can be viewed as navigating deep attractor basins, leading to a 5.38% performance boost on GSM8K when leveraging Gibbs-weighted retrieval.

Extending the prediction horizon of chaotic systems by 2.3x reveals that chaos isn't inherently unpredictable—it's a matter of the right neural architecture.