List your company

Long-Context Generation on PG-19 (50K context length)

5.79Throughput Speedup (micro)SpecPV-2K
1.91082.91793.9254.9321Dec 2, 2025

Evaluation results

MethodLinks
SpecPV-2KBackbone=LLaMA3.1-8B-Instruct, Partial KV Budget=2K2025.125.793.56
SpecPV-4KBackbone=LLaMA3.1-8B-Instruct, Partial KV Budget=4K2025.125.663.52
SpecPV-8KBackbone=LLaMA3.1-8B-Instruct, Partial KV Budget=8K2025.125.313.46
TriForceBackbone=LLaMA3.1-8B-Instruct, Partial KV Budget=4K2025.123.422.82
EAGLE3-YARNBackbone=LLaMA3.1-8B-Instruct2025.123.063.46
TokenSwiftBackbone=LLaMA3.1-8B-Instruct, Partial KV Budget=4K2025.122.061.77