ByteDance-Seed

AHN: Artificial Hippocampus Networks for Efficient Long-Context Modeling

Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference

WideSearch: Benchmarking Agentic Broad Info-Seeking

[Arxiv 2025] ByteMorph: Benchmarking Instruction-Guided Image Editing with Non-Rigid Motions