arXiv:2606. 20945v2 Announce Type: replace Abstract: Self-attention is central to Transformer performance and is often the most expensive part of the Transformer at long context lengths because its pairwise token interactions scale quadratically with sequence length.
arXiv:2609.13285v1 Announce Type: cross Abstract: The KV cache is a primary bottleneck for Transformer decoding: its memory footprint and cache-read traffic grow with sequence length. Grouped-query a...
arXiv:2606. 21848v2 Announce Type: replace-cross Abstract: Transformer architectures form the foundation of modern natural language processing, making it crucial to address the efficiency and scalability limitations of the standard QKV attention mechanism.
arXiv:2608. 11519v1 Announce Type: new Abstract: Full self-attention is a strong token mixer for PDE surrogates on irregular domains, but its quadratic cost limits its use on high-resolution problems.
Attention-Aware Routing (AAR) augments the router in Mixture-of-Experts language models with temporal and spectral features derived from a sliding window of attention weights, thereby separating contextual information from the token’s hidden state. By keeping the base transformer frozen and training only routing parameters, AAR achieves a +3.37‑point improvement on GSM8K over a routing‑only baseline and demonstrates that routing changes propagate through the residual stream to reshape attention without directly updating the attention mechanism. The method also reduces long diverging generations, shows depth‑sensitivity affecting retrieval versus reasoning, and offers a controlled probe of routing‑relevant information across layers.
arXiv:2607. 06601v1 Announce Type: cross Abstract: Conditional computation can decouple language model quality from per-token inference cost, yet leading techniques act on a single axis in isolation: Mixture-of-Experts (MoE) sparsifies the FFN, Mixture-of-Depths (MoD) skips whole transformer blocks, and KV-cache quantization compresses attention memory.
arXiv:2606. 06467v1 Announce Type: cross Abstract: Long-context inference in modern LLMs is increasingly constrained by decoding efficiency, especially in reasoning-heavy settings where models generate long intermediate chains of thought.
arXiv:2608. 08853v1 Announce Type: new Abstract: Sparse Mixture-of-Experts (MoE) routers commonly use the same scores both to select experts and to weight their already-computed outputs.
The paper examines why learned gates in sparse attention models offer little advantage over random gates when jointly trained with the transformer. Through experiments on a 31M-parameter transformer, the authors attribute this to routing absorption, where the model’s representations adapt to the imposed mask, diminishing the benefit of learned routing. They also explore hard masking, stochastic mask training, and the impact of trainable attention layers on gate performance, concluding that freezing the model stabilizes routing targets for effective post‑hoc sparsification.
The paper introduces Ban&Pick, a post‑training, plug‑and‑play routing strategy for Sparse Mixture‑of‑Experts large language models. It identifies and reinforces a small group of highly influential experts while dynamically pruning redundant ones, leading to accuracy gains across math, code, and reasoning benchmarks. Experiments on DeepSeek and Qwen3 show notable performance improvements and a 1.25× inference speedup without retraining or architectural changes.
The paper introduces ReadKV, a query‑adaptive quantization method for KV caches that stores keys and values in progressive codes, allowing each decoding query to allocate different precision prefixes for keys and values. By optimizing a calibrated distortion objective under fixed budgets, ReadKV achieves near‑optimal allocation and demonstrates that query‑dependent access can outperform any query‑independent reader at the same read budget. Experiments on six base models show that reading only four bits on average from an eight‑bit cache increases perplexity by at most 0.66% while reducing logical reads and retained capacity compared to a 16‑bit cache, and it achieves lower latency than existing codecs on an NVIDIA A10G.
arXiv:2608. 03276v1 Announce Type: new Abstract: Long-context inference with large language models is constrained by the linear growth of the key-value cache to sequence length.
The paper introduces Post-Boundary Bridge (PBB), a hybrid transformer architecture that keeps causal attention within blocks while adding direct connections across block boundaries. PBB focuses on within-block modeling and nearby exchange, delegating long-range communication to full-attention layers. Experiments on dense and mixture-of-experts models ranging from 205 million to 2.07 billion parameters show that PBB hybrids maintain near-full perplexity, competitive downstream performance, and improved source retrieval, while Flash-PBB achieves 1.82× faster decoding with half the local key‑value cache compared to Flash‑SWA.