All news
4 articles- paperHugging Face Daily Papers·20 Oct 20264.4Autonomous AI research teams discover linear attention architectures.
- paperEmergent Mind·25 Oct 20264.2Kronecker Delta Rule treats key-value pairs as joint addresses in linear attention to reduce interference at long context.
- paperHugging Face Daily Papers·21 Oct 20264.1PACE encodes linear-attention recurrent states in four bits.
- paperarXiv (systems, data, ML)·23 Oct 20264.0Recasts sequence compression as test-time regression for efficient models.