List your company

Long Reasoning on MATH500 (pass@k, avg@k)

78Pass@k RateFull
67.670.37375.7May 19, 2025

Evaluation results

MethodLinks
FullBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=Full2025.057867.25
FreeKVBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=FreeKV2025.057866.75
FullBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=Full2025.057871.75
FreeKVBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=FreeKV2025.057870
FreeKVBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=FreeKV2025.057867.5
ShadowKVBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=ShadowKV2025.057660.25
QuestBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=Quest2025.057668
ArkValeBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=ArkVale2025.057668.25
QuestBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=Quest2025.057667.25
ShadowKVBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=ShadowKV2025.057665
RaaSBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=RaaS2025.057462.5
InfiniGenBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=InfiniGen2025.057466.5
RaaSBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=RaaS2025.057467
ShadowKVBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=ShadowKV2025.057464.75
InfiniGenBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=InfiniGen2025.057470
FullBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=Full2025.057470.25
RazorBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=Razor2025.057260.5
QuestBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=Quest2025.057262
ArkValeBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=ArkVale2025.057262.5
RazorBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=Razor2025.057266.75
ArkValeBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=ArkVale2025.057266.25
InfiniGenBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=InfiniGen2025.057267
RazorBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=Razor2025.057059.75
RaaSBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=RaaS2025.056864.75