Search posts, papers, and topics
Fathom introduces a per-query read depth mechanism for sparse decoding over offloaded KV caches, allowing each query to adaptively decide how many bits of each key channel to read. This method significantly speeds up decoding for large language models with long contexts by reducing host memory traffic, achieving 1.67x faster GPU decoding on Qwen3-8B at one million tokens.
This paper introduces SpectralShift, a spectral reparameterization method for extending the context window of Gated DeltaNet (GDN) linear attention models. It reconfigures the decay spectrum by enhancing slow propagation and preserving fast-decaying modes, consistently improving long-context capabilities during continual pretraining.
This paper investigates the "lossless" claim of Orthrus, a hybrid architecture for accelerating LLM inference. It finds that under BF16 precision, Orthrus diverges from the exact autoregressive output trajectory in over 50% of cases, though FP32 maintains exact matching. Despite BF16 divergence, downstream task performance was not systematically degraded.
