List your company

Question Answering on TriviaQA (Accuracy, Rejection Rate)

62AccuracyFaster Cascades
48.4851.9955.559.01Jun 10, 2026

Evaluation results

MethodLinks
Faster CascadesModel=Qwen 7B→72B, Method=Faster Cascades, Speedup=2.25×2026.066221
VIA-SDModel=Qwen 7B→72B, Method=VIA-SD (ours), Speedup=2.55×2026.066117
Speculative DecodingModel=Qwen 7B→72B, Method=Speculative Decoding, Speedup=1.70×2026.066026
Cascade SDModel=Qwen 7B→72B, Method=Cascade SD, Speedup=1.90×2026.066025
VIA-SDModel=Qwen 7B→14B, Method=VIA-SD (ours), Speedup=2.40×2026.065715
Faster CascadesModel=Gemma2 2B→27B, Method=Faster Cascades, Speedup=2.30×2026.065620
Faster CascadesModel=LLaMA2 7B→70B, Method=Faster Cascades, Speedup=2.15×2026.065622
Faster CascadesModel=Qwen 7B→14B, Method=Faster Cascades, Speedup=1.95×2026.065619
VIA-SDModel=Gemma2 2B→27B, Method=VIA-SD (ours), Speedup=2.50×2026.065515
VIA-SDModel=LLaMA2 7B→70B, Method=VIA-SD (ours), Speedup=2.45×2026.065518
Speculative DecodingModel=Gemma2 2B→27B, Method=Speculative Decoding, Speedup=1.65×2026.065424
Cascade SDModel=Gemma2 2B→27B, Method=Cascade SD, Speedup=1.81×2026.065423
Speculative DecodingModel=LLaMA2 7B→70B, Method=Speculative Decoding, Speedup=1.60×2026.065427
Cascade SDModel=LLaMA2 7B→70B, Method=Cascade SD, Speedup=1.80×2026.065426
Speculative DecodingModel=Qwen 7B→14B, Method=Speculative Decoding, Speedup=1.58×2026.065424
Cascade SDModel=Qwen 7B→14B, Method=Cascade SD, Speedup=1.70×2026.065423
VIA-SDModel=Gemma2 2B→9B, Method=VIA-SD (ours), Speedup=2.33×2026.065313
Faster CascadesModel=Gemma2 2B→9B, Method=Faster Cascades, Speedup=1.90×2026.065217
VIA-SDModel=LLaMA2 7B→13B, Method=VIA-SD (ours), Speedup=2.20×2026.065216
Faster CascadesModel=LLaMA2 7B→13B, Method=Faster Cascades, Speedup=1.82×2026.065120
Speculative DecodingModel=Gemma2 2B→9B, Method=Speculative Decoding, Speedup=1.55×2026.065021
Cascade SDModel=Gemma2 2B→9B, Method=Cascade SD, Speedup=1.62×2026.065020
Speculative DecodingModel=LLaMA2 7B→13B, Method=Speculative Decoding, Speedup=1.48×2026.064926
Cascade SDModel=LLaMA2 7B→13B, Method=Cascade SD, Speedup=1.60×2026.064924