List your company

Question Answering on TruthfulQA (Factuality and Diversity Metrics)

29.62FactualityTypical + DFD (MPT-7B)
24.024825.477426.9328.3826Mar 11, 2025

Evaluation results

MethodLinks
Typical + DFD (MPT-7B)Backbone=MPT-7B, Decoding Strategy=Typical, Integration of DFD=True2025.0329.6276.3584.4584.1513.98
Top-k + DFD (MPT-7B)Backbone=MPT-7B, Decoding Strategy=Top-k, Integration of DFD=True2025.0328.476.0484.9584.813.37
Nucleus + DFD (MPT-7B)Backbone=MPT-7B, Decoding Strategy=Nucleus, Integration of DFD=True2025.0328.475.6683.8283.6114.75
Top-k (MPT-7B)Backbone=MPT-7B, Decoding Strategy=Top-k, Integration of DFD=False2025.0327.2574.5283.5983.8614.5
Nucleus (MPT-7B)Backbone=MPT-7B, Decoding Strategy=Nucleus, Integration of DFD=False2025.0326.6874.8282.6182.3215.88
Temperature + DFD (MPT-7B)Backbone=MPT-7B, Decoding Strategy=Temperature, Integration of DFD=True2025.0325.8380.7289.0288.028.84
Typical (MPT-7B)Backbone=MPT-7B, Decoding Strategy=Typical, Integration of DFD=False2025.0325.775.7783.3883.0415.29
Temperature (MPT-7B)Backbone=MPT-7B, Decoding Strategy=Temperature, Integration of DFD=False2025.0324.2479.9688.587.689.26