The paper introduces the Holographic Quantum Transformer (HQT), a generative architecture using global self-attention to model non-local entanglement in two-dimensional frustrated quantum systems. On the 8×8 square-lattice J1-J2 Heisenberg model at the critical point J2=0.5, it reports a ground-state energy per site of -0.5001(1). Its “Holographic Transfer” protocol projects a model trained on 8×8 systems to 10×10 lattices using continuous positional-embedding interpolation and attention-head re-initialization, reaching -0.49782(3) without training from scratch.
No heat snapshots are available in the last 24 hours.