List your company

Long-context understanding on LongBench v1 (test)

49.6SD QALlama-3.1-8B
18.81626.80834.842.792May 26, 2025

Evaluation results

MethodLinks
Llama-3.1-8BBase Model=Llama-3.1-8B, Tokens=128k2025.0549.650.931.269.46053.553.8
TailorKV-2Base Model=Llama-3.1-8B, Tokens=64(+128)2025.0549.350.529.468.758.153.352.9
PQCacheBase Model=Llama-3.1-8B, Tokens=1922025.0548.449.52767.356.353.651.7
TailorKV-1Base Model=Llama-3.1-8B, Tokens=64(+128)2025.0548.250.929.268.158.353.452.6
QuestBase Model=Llama-3.1-8B, Tokens=1922025.0540.146.920.761.64852.446.2
TailorKV-1Base Model=Yi-9B, Tokens=64(+128)2025.053844.327.360.266.324.344.7
Yi-9BBase Model=Yi-9B, Tokens=200k2025.0536.644.728.860.669.63547
TailorKV-2Base Model=Yi-9B, Tokens=64(+128)2025.0535.643.527.360.16623.544
SnapKVBase Model=Llama-3.1-8B, Tokens=1922025.0535.248.120.256.552.852.545.2
TailorKV-1Base Model=Yi-6B, Tokens=64(+128)2025.0532.515.41.449.755.9428.3
TailorKV-2Base Model=Yi-6B, Tokens=64(+128)2025.0532.515.31.549.156.4428.2
PQCacheBase Model=Yi-9B, Tokens=1922025.0532.441.619.258.664.427.842
Yi-6BBase Model=Yi-6B, Tokens=200k2025.0532.415.31.349.969.89.529.7
PQCacheBase Model=Yi-6B, Tokens=1922025.0530.414.50.64855.8427.3
QuestBase Model=Yi-9B, Tokens=1922025.0529.237.915.457.559.625.739.1
QuestBase Model=Yi-6B, Tokens=1922025.0526.512.50.346.951.98.526.2
StreamLLMBase Model=Llama-3.1-8B, Tokens=1922025.0526.342.717.95048.253.540.6
SnapKVBase Model=Yi-9B, Tokens=1922025.052538.811.94959.718.835
SnapKVBase Model=Yi-6B, Tokens=1922025.0524.2131.638.551.23.723.3
StreamLLMBase Model=Yi-9B, Tokens=1922025.0521.333.61144.151.814.730.6
StreamLLMBase Model=Yi-6B, Tokens=1922025.052011.61.63444.6420.4