List your company
ResearchFormerly WizWand

Best feasible prompt identification on XSum

0.161Average Soft Constrained Reward (b=3)CSR
0.009160.048580.0880.12742May 14, 2026
Updated 6mo ago

Evaluation results

MethodLinks
CSRBackbone Model=Llama3-8B, K (Prompt pool size)=302026.050.1610.1540.1480.149
CSRBackbone Model=Llama3-8B, K (Prompt pool size)=502026.050.1440.1340.1220.157
MLP-CSRBackbone Model=Gemma-7B, K (Prompt pool size)=502026.050.1430.1420.1470.142
MLP-CSRBackbone Model=Llama3-8B, K (Prompt pool size)=302026.050.1430.1260.1470.149
MLP-CSRBackbone Model=Llama3-8B, K (Prompt pool size)=502026.050.1410.1510.1540.156
CSRBackbone Model=Llama3-8B, K (Prompt pool size)=1002026.050.1410.1440.1530.154
MLP-CSRBackbone Model=Gemma-7B, K (Prompt pool size)=1002026.050.1390.1420.1450.147
CSRBackbone Model=Gemma-7B, K (Prompt pool size)=1002026.050.1340.1410.1420.143
MLP-CSRBackbone Model=Llama3-8B, K (Prompt pool size)=1002026.050.1240.1440.1580.157
MLP-CSRBackbone Model=Gemma-7B, K (Prompt pool size)=302026.050.1230.1390.140.142
CSRBackbone Model=Gemma-7B, K (Prompt pool size)=502026.050.1220.1430.1410.14
CSRBackbone Model=Gemma-7B, K (Prompt pool size)=302026.050.1170.1370.1440.143
UniformBackbone Model=Llama3-8B, K (Prompt pool size)=1002026.050.1130.1340.1230.141
UniformBackbone Model=Llama3-8B, K (Prompt pool size)=502026.050.0630.0450.0510.078
UniformBackbone Model=Llama3-8B, K (Prompt pool size)=302026.050.0480.0690.080.089
UniformBackbone Model=Gemma-7B, K (Prompt pool size)=502026.050.0210.0210.0370.036
UniformBackbone Model=Gemma-7B, K (Prompt pool size)=1002026.050.0210.0660.0370.044
UniformBackbone Model=Gemma-7B, K (Prompt pool size)=302026.050.01500.030.037