List your company

Open-book generation under knowledge conflict on ConFiQA 1,500 subset

81.07Ps ScoreContextDPO
9.72628.24846.7765.292Jan 7, 2026

Evaluation results

MethodLinks
ContextDPOModel=LLAMA-3.1-8B, O&I Prompting=true2026.0181.078.879.86
ContextFocusModel=LLAMA-3.1-8B, O&I Prompting=true, Steering Layer=13, Steering Multiplier (m)=22026.0177.537.68.93
ContextDPOModel=MISTRAL-7B-INSTRUCT, O&I Prompting=true2026.0177.410.1311.58
ContextFocusModel=MISTRAL-7B-INSTRUCT, O&I Prompting=true, Steering Layer=11, Steering Multiplier (m)=22026.0175.1311.5313.31
ContextDPOModel=MISTRAL-7B-INSTRUCT, O&I Prompting=false2026.0172.9313.3315.46
COIECDModel=MISTRAL-7B-INSTRUCT, O&I Prompting=true2026.017113.9316.41
ContextFocusModel=LLAMA-3.1-8B, O&I Prompting=false, Steering Layer=13, Steering Multiplier (m)=22026.0170.879.2711.56
BaseModel=LLAMA-3.1-8B, O&I Prompting=true2026.0170.415.417.95
COIECDModel=LLAMA-3.1-8B, O&I Prompting=true2026.0169.6714.2717
ContextFocusModel=MISTRAL-7B-INSTRUCT, O&I Prompting=false, Steering Layer=11, Steering Multiplier (m)=22026.0167.816.7319.79
BaseModel=MISTRAL-7B-INSTRUCT, O&I Prompting=true2026.0166.2715.5318.99
COIECDModel=MISTRAL-7B-INSTRUCT, O&I Prompting=false2026.0162.1321.8726.03
ContextDPOModel=LLAMA-3.1-8B, O&I Prompting=true2026.0159.1317.3322.67
BaseModel=MISTRAL-7B-INSTRUCT, O&I Prompting=false2026.0158.9322.8727.95
ContextDPOModel=LLAMA-3.1-8B, O&I Prompting=false2026.0158.418.624.16
ContextDPOModel=MISTRAL-7B-INSTRUCT, O&I Prompting=true2026.0156.9317.223.2
ContextFocusModel=LLAMA-3.1-8B, O&I Prompting=false, Steering Layer=13, Steering Multiplier (m)=22026.0153.3310.216.05
ContextFocusModel=MISTRAL-7B-INSTRUCT, O&I Prompting=true, Steering Layer=11, Steering Multiplier (m)=22026.0150.615.5323.49
ContextFocusModel=LLAMA-3.1-8B, O&I Prompting=true, Steering Layer=13, Steering Multiplier (m)=22026.0148.610.2717.44
ContextDPOModel=LLAMA-3.1-8B, O&I Prompting=false2026.0146.2717.227.1
COIECDModel=MISTRAL-7B-INSTRUCT, O&I Prompting=true2026.0142.4716.427.86
BaseModel=MISTRAL-7B-INSTRUCT, O&I Prompting=true2026.0137.0716.6731.02
COIECDModel=LLAMA-3.1-8B, O&I Prompting=true2026.0136.617.5332.39
COIECDModel=LLAMA-3.1-8B, O&I Prompting=false2026.0135.4732.647.89
BaseModel=LLAMA-3.1-8B, O&I Prompting=false2026.0135.2732.3347.83
ContextFocusModel=MISTRAL-7B-INSTRUCT, O&I Prompting=false, Steering Layer=11, Steering Multiplier (m)=22026.0135.1319.3335.5
ContextDPOModel=MISTRAL-7B-INSTRUCT, O&I Prompting=false2026.0132.8718.4735.97
BaseModel=LLAMA-3.1-8B, O&I Prompting=true2026.013018.5338.19
COIECDModel=MISTRAL-7B-INSTRUCT, O&I Prompting=false2026.0125.9318.441.5
BaseModel=MISTRAL-7B-INSTRUCT, O&I Prompting=false2026.0119.6719.3349.57
COIECDModel=LLAMA-3.1-8B, O&I Prompting=false2026.0114.819.857.23
BaseModel=LLAMA-3.1-8B, O&I Prompting=false2026.0112.4717.257.98