List your company

Long-context understanding on LongBench v2 (test)

35.2Overall ScoreSoLo-DPO
24.17627.03829.932.762May 16, 2025

Evaluation results

MethodLinks
SoLo-DPOBackbone=Qwen2.5-7B-Instruct, Mode=SoLo2025.0535.237.533.839.331.835
SoLo-ORPOBackbone=Llama3.1-8B-Instruct, Mode=SoLo2025.0534.737.5334032.131.2
ORPOBackbone=Llama3.1-8B-Instruct, Mode=short2025.0533.736.831.841.229.828.7
LongPO(pub)Backbone=Qwen2.5-7B-Instruct2025.0533.3353240.53027.8
SoLo-ORPOBackbone=Qwen2.5-7B-Instruct, Mode=SoLo2025.0533.236.331.239.728.830.9
InstructBackbone=Llama3.1-8B-Instruct2025.0532.535.530.740.727.728.5
ORPOBackbone=Llama3.1-8B-Instruct, Mode=long2025.0532.134.530.639.727.828
SFTBackbone=Llama3.1-8B-Instruct, Mode=short2025.0531.834.53038.328.627
SoLo-SimPOBackbone=Qwen2.5-7B-Instruct, Mode=SoLo2025.053134.129.137.525.730.6
SFTBackbone=Llama3.1-8B-Instruct, Mode=long2025.0530.936.827.33629.624.8
SFTBackbone=Qwen2.5-7B-Instruct, Mode=short2025.0530.833.629.139.225.727.2
SFTBackbone=Qwen2.5-7B-Instruct, Mode=long2025.05303228.835.725.928.7
DPOBackbone=Qwen2.5-7B-Instruct, Mode=long2025.0529.734.326.935.925.627.6
LongPO(reimp)Backbone=Qwen2.5-7B-Instruct2025.0529.632.22836.726.723.7
InstructBackbone=Qwen2.5-7B-Instruct2025.0529.330.928.336.924.626.1
ORPOBackbone=Qwen2.5-7B-Instruct, Mode=short2025.0529.334.426.23525.327.8
ORPOBackbone=Qwen2.5-7B-Instruct, Mode=long2025.0526.630.124.533.822.323.3
DPOBackbone=Qwen2.5-7B-Instruct, Mode=short2025.0525.527.224.529.723.522.8
SimPOBackbone=Qwen2.5-7B-Instruct, Mode=long2025.0525.425.425.433—23.3
SimPOBackbone=Qwen2.5-7B-Instruct, Mode=short2025.0524.627.123.129.521.223.1