List your company

Single-turn data-to-logic tasks on DTLBench

90.29DDXPlus ScoreCASCADE
34.431648.933363.43577.9367May 5, 2026

Evaluation results

MethodLinks
CASCADEBackbone=gemini-2.0-flash2026.0590.29———67.1361.9983.3340.7975.9686.2983.0564.4272.58
NP-CBRBackbone=gemini-2.0-flash2026.0586.91———66.3362.0581.9437.6469.0285.1182.8564.370.68
CASCADEBackbone=Qwen3-32B2026.0584.5463.9856.7952.2562.4460.2684.6647.7270.7686.1575.4155.1866.68
NP-CBRBackbone=Qwen3-32B2026.0580.8460.5856.2951.8360.9359.5682.7745.3565.0885.3966.8749.663.76
REINFORCE+LoRABackbone=Qwen3-32B2026.0576.913.8857.4753.9760.3252.6373.8747.2658.286.873.5452.3558.93
CASCADEBackbone=Qwen3-14B2026.0575.6662.3557.7654.5556.9559.0682.1645.173.6885.6174.6653.6865.1
ICRLBackbone=gemini-2.0-flash2026.0573.79———64.9855.7774.130.0728.584.780.6262.3261.65
ICRLPlusBackbone=gemini-2.0-flash2026.0573.52———6656.5573.9231.8230.0984.680.6161.9962.12
REINFORCE+LoRABackbone=Qwen3-14B2026.0571.311.0758.2152.3953.7241.3872.344.4254.1186.4772.0749.1755.55
NP-CBRBackbone=Qwen3-14B2026.0570.6961.2856.7753.1556.3356.1181.0242.9866.4185.3171.5752.1462.81
CASCADEBackbone=Qwen3-4B2026.0570.04—57.4336.3346.8749.1280.137.9657.4586.4575.954.5959.29
Zero-shotBackbone=gemini-2.0-flash2026.0569.67———59.6139.3273.6112.9228.0683.1680.6662.1956.58
CASCADEBackbone=Qwen3-8B2026.0569.1354.8863.9751.247.947.7580.7839.7766.9186.472.748.6660.84
ICRLBackbone=Qwen3-32B2026.0566.1112.9550.8250.2357.0344.0667.5933.7741.8383.5570.9550.8652.48
REINFORCE+LoRABackbone=Qwen3-8B2026.0565.554.9464.1652.0448.8551.6768.8443.1954.1786.3769.0844.7754.47
NP-CBRBackbone=Qwen3-8B2026.0565.4452.8562.7349.9145.4546.8678.9238.0460.9385.366.7448.1158.44
ICRLPlusBackbone=Qwen3-32B2026.0564.9130.2952.2451.0558.3343.2267.6135.3943.3483.9171.1150.5254.33
Zero-shotBackbone=Qwen3-32B2026.0563.7712.7951.3648.5625.5339.1266.6634.345.3684.0563.8444.6548.33
NP-CBRBackbone=Qwen3-4B2026.0562.69—56.5634.9344.8445.9277.3635.1953.785.4273.4352.9356.63
REINFORCE+LoRABackbone=Qwen3-4B2026.0561.39—59.6544.747.6945.8965.8935.6649.3185.5773.4253.1256.57
ICRLBackbone=Qwen3-14B2026.0556.51.1854.6548.4749.3940.5266.9932.4432.0382.1269.2948.5948.51
ICRLPlusBackbone=Qwen3-14B2026.0556.1635.3955.3949.750.0941.9167.1434.7138.4782.5269.3948.3352.43
Zero-shotBackbone=Qwen3-14B2026.0555.7511.9152.9240.9250.8438.6565.8633.0336.6678.4166.2948.7548.33
ICRLBackbone=Qwen3-8B2026.0549.53.1953.446.7640.6829.460.4326.1221.1181.5367.6643.2543.59
ICRLPlusBackbone=Qwen3-8B2026.0548.5723.3655.2348.7542.4831.6960.4527.9728.5580.9868.1944.8646.76
Zero-shotBackbone=Qwen3-8B2026.0545.25.9450.4935.639.7636.2261.3726.4938.0874.6160.9643.9643.22
ICRLBackbone=Qwen3-4B2026.0544.52—51.0128.4335.921.662.9227.1521.7283.3171.6951.7145.45
ICRLPlusBackbone=Qwen3-4B2026.0544.11—50.131.6938.8927.8863.0427.9926.0583.3771.6752.8647.06
Zero-shotBackbone=Qwen3-4B2026.0536.58—50.5326.8534.7219.2462.3324.7223.5281.9270.1552.0743.88