BWhat changed
Used dilated attention with exponentially growing receptive fields and sequence-partitioned execution to reduce long-sequence attention computation and communication.
Scientific scope
Used dilated attention with exponentially growing receptive fields and sequence-partitioned execution to reduce long-sequence attention computation and communication. Billion-token evidence is distributed forward-runtime feasibility, not a trained model demonstrating billion-token reasoning or end-to-end learning quality. Sparse attention is not exact dense attention; scaling depends on the stated device/data layout.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Billion-token evidence is distributed forward-runtime feasibility, not a trained model demonstrating billion-token reasoning or end-to-end learning quality. Sparse attention is not exact dense attention; scaling depends on the stated device/data layout.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
- LongNet: Scaling Transformers to 1,000,000,000 Tokens
- Ding et al.: LongNet dilated distributed attention — official project source
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction