List your company

Mathematical Reasoning on GSM8k (flexible_extract)

49.96Flexible Extract ScoreBaseline
-1.5211.84525.2138.575Feb 26, 2026

Evaluation results

MethodLinks
BaselineModel=Llama 3.1-8B2026.0249.96
TurboQuantModel=Llama 3.1-8B2026.0248.9
InnerQBaseModel=Llama 3.1-8B2026.0248.22
KIVISinkModel=Llama 3.1-8B2026.0247.84
InnerQHybridModel=Llama 3.1-8B2026.0247.31
InnerQSmallModel=Llama 3.1-8B2026.0245.94
KIVIModel=Llama 3.1-8B2026.0242.84
BaselineModel=Mistral 7B v0.32026.0236.92
InnerQBaseModel=Mistral 7B v0.32026.0236.47
TurboQuantModel=Mistral 7B v0.32026.0234.72
InnerQHybridModel=Mistral 7B v0.32026.0233.89
InnerQSmallModel=Mistral 7B v0.32026.0233.89
KIVISinkModel=Mistral 7B v0.32026.0233.81
KIVIModel=Mistral 7B v0.32026.0232.3
InnerQBaseModel=Llama 3.2-3B2026.0227.14
BaselineModel=Llama 3.2-3B2026.0226.46
TurboQuantModel=Llama 3.2-3B2026.0225.47
InnerQSmallModel=Llama 3.2-3B2026.0224.03
KIVISinkModel=Llama 3.2-3B2026.0223.35
InnerQHybridModel=Llama 3.2-3B2026.0223.12
KIVIModel=Llama 3.2-3B2026.0219.33
InnerQBaseModel=Llama 2-7B2026.0213.42
KIVIModel=Llama 2-7B2026.0213.12
BaselineModel=Llama 2-7B2026.0213.04
TurboQuantModel=Llama 2-7B2026.0212.59
KIVISinkModel=Llama 2-7B2026.0211.9
InnerQHybridModel=Llama 2-7B2026.0211.3
BaselineModel=Llama 3.2-1B2026.026.22
TurboQuantModel=Llama 3.2-1B2026.025.84
InnerQBaseModel=Llama 3.2-1B2026.025.53
InnerQHybridModel=Llama 3.2-1B2026.025.31
InnerQSmallModel=Llama 3.2-1B2026.024.63
KIVISinkModel=Llama 3.2-1B2026.024.02
KIVIModel=Llama 3.2-1B2026.022.81
InnerQSmallModel=Llama 2-7B2026.020.46