List your company

Factuality Hallucination Evaluation on LongFact (test)

100Response ScoreVanilla
9597.5100102.5Feb 11, 2025

Evaluation results

MethodLinks
VanillaInstruction Dataset=Alpaca, Sample Ratio=100%2025.0210017.185.880.383.1—————
FLAME-DPO factInstruction Dataset=Alpaca, Sample Ratio=100%2025.0210015.688.381.284.8—————
SELF-EVALInstruction Dataset=Alpaca, Sample Ratio=100%2025.0210016.987.88184.4—————
IFDInstruction Dataset=Alpaca, Sample Ratio=5%2025.021002187.280.583.9—————
CaRInstruction Dataset=Alpaca, Sample Ratio=5%2025.0210016.286.679.182.9—————
NuggetsInstruction Dataset=Alpaca, Sample Ratio=5%2025.0210018.388.681.284.9—————
NOVAInstruction Dataset=Alpaca, Sample Ratio=5%2025.0210017.992.482.787.6—————
IFDInstruction Dataset=Alpaca, Sample Ratio=10%2025.0210020.586.379.282.8—————
CaRInstruction Dataset=Alpaca, Sample Ratio=10%2025.0210024.387.181.384.2—————
NuggetsInstruction Dataset=Alpaca, Sample Ratio=10%2025.0210027.186.780.483.6—————
NOVAInstruction Dataset=Alpaca, Sample Ratio=10%2025.0210018.489.181.685.4—————
IFDInstruction Dataset=Alpaca, Sample Ratio=15%2025.0210019.484.780.782.7—————
CaRInstruction Dataset=Alpaca, Sample Ratio=15%2025.0210020.986.47882.2—————
NuggetsInstruction Dataset=Alpaca, Sample Ratio=15%2025.0210023.486.580.183.3—————
NOVAInstruction Dataset=Alpaca, Sample Ratio=15%2025.0210018.788.182.185.1—————
VanillaInstruction Dataset=Alpaca - GPT4, Sample Ratio=100%2025.021003284.780.482.6—————
FLAME-DPO factInstruction Dataset=Alpaca - GPT4, Sample Ratio=100%2025.0210027.687.384.185.7—————
SELF-EVALInstruction Dataset=Alpaca - GPT4, Sample Ratio=100%2025.0210031.686.783.785.2—————
IFDInstruction Dataset=Alpaca - GPT4, Sample Ratio=5%2025.0210039.284.479.682—————
CaRInstruction Dataset=Alpaca - GPT4, Sample Ratio=5%2025.0210041.186.281.183.7—————
NuggetsInstruction Dataset=Alpaca - GPT4, Sample Ratio=5%2025.0210042.38782.384.7—————
NOVAInstruction Dataset=Alpaca - GPT4, Sample Ratio=5%2025.0210033.890.185.287.7—————
IFDInstruction Dataset=Alpaca - GPT4, Sample Ratio=10%2025.0210039.286.577.882.2—————
CaRInstruction Dataset=Alpaca - GPT4, Sample Ratio=10%2025.021003887.178.382.7—————
NuggetsInstruction Dataset=Alpaca - GPT4, Sample Ratio=10%2025.0210035.788.280.184.2—————
NOVAInstruction Dataset=Alpaca - GPT4, Sample Ratio=10%2025.0210032.390.681.886.2—————
IFDInstruction Dataset=Alpaca - GPT4, Sample Ratio=15%2025.0210032.285.280.382.8—————
CaRInstruction Dataset=Alpaca - GPT4, Sample Ratio=15%2025.0210033.685.881.583.7—————
NuggetsInstruction Dataset=Alpaca - GPT4, Sample Ratio=15%2025.0210034.586.180.783.4—————
NOVAInstruction Dataset=Alpaca - GPT4, Sample Ratio=15%2025.0210032.18882.585.3—————
Entropy2025.12—————0.512410.76———
Linear Probe2025.12—————0.902268.9190.2288.2952.8
MLP Probe2025.12—————0.952870.2495.2889.0257.15
Perplexity2025.12—————0.46238.79———