Dustin: Draft-Augmented Sparse Verification for Efficient Long-Context Generation with Speculative Decoding
About
While speculative decoding improves inference throughput for multi-batch long-context Large Language Models (LLMs), its efficiency is often limited by a verification bottleneck where Key-Value (KV) cache loading dominates latency. Existing compression methods fail in this regime: static eviction incurs accuracy loss due to saliency shift, while dynamic selection introduces prohibitive computational overhead during the verification path. We propose Dustin, a sparse verification framework designed for long-context speculative decoding. Dustin integrates lookahead signals from the draft model with historical attention from the target model to identify critical tokens with high fidelity across multi-step verification windows. To reduce recomputation latency, this approach further employs a sparse estimation scheme that restricts importance scoring to a minimal subset of attention heads. Evaluations on PG-19 and LongBench with Qwen2.5-72B demonstrate that Dustin achieves a 27.85x speedup in self-attention and a 9.17x end-to-end decoding speedup at a 32k sequence length, all with negligible accuracy degradation.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Long-context language modeling | LongBench | -- | 369 | |
| Long-context Understanding | LongBench (test) | Avg Score51.46 | 166 | |
| Long-context Understanding | LongBench | Overall Average Score55.23 | 143 | |
| Long-Context Generation Efficiency | PG-19 | Throughput408 | 64 | |
| Long-context generation | PG-19 8K context length | Throughput553.8 | 32 | |
| Long-context generation | PG-19 16K context length | Throughput486.5 | 32 | |
| Long-context generation | PG-19 24K context length | Throughput393.9 | 32 | |
| Long-context generation | PG-19 32K context length | Throughput329.6 | 32 | |
| Long-Context Generation Efficiency | Qwen2.5-72B-Instruct Long-Context Efficiency Workload | Speedup (8K)4.76 | 2 |