List your company
ResearchFormerly WizWand

Truthfulness Evaluation on TruthfulQA (Avg@k, Pass@k)

66.63Avg@kDFT
49.52253.963558.40562.8465May 8, 2026
Updated 5mo ago

Evaluation results

MethodLinks
DFTBackbone=Qwen2.5-7B-Instruct2026.0566.6379.68
DFTBase Model=Qwen2.5-7B-Instruct, RL Stage=DAPO, k=42026.0566.6379.68
BaseBackbone=Qwen2.5-7B-Instruct2026.0566.3778.51
SFTBackbone=Qwen2.5-7B-Instruct2026.0562.4380.26
SFTBase Model=Qwen2.5-7B-Instruct, RL Stage=DAPO, k=42026.0562.4380.26
IWBackbone=Qwen2.5-7B-Instruct2026.0561.4878.07
IWBase Model=Qwen2.5-7B-Instruct, RL Stage=DAPO, k=42026.0561.4878.07
RPSFTBackbone=Qwen2.5-7B-Instruct2026.0561.478.36
RPSFTBase Model=Qwen2.5-7B-Instruct, RL Stage=DAPO, k=42026.0561.478.36
BaseBackbone=Qwen2.5-3B-Instruct2026.0561.0471.78
DFTBackbone=Llama-3.1-8B-Instruct2026.0560.5669.15
DFTBase Model=Llama-3.1-8B-Instruct, RL Stage=DAPO, k=42026.0560.5669.15
BaseBackbone=Llama-3.1-8B-Instruct2026.0559.579.82
DFTBackbone=Qwen2.5-3B-Instruct2026.0557.9373.83
DFTBase Model=Qwen2.5-3B-Instruct, RL Stage=DAPO, k=42026.0557.9373.83
RPSFTBackbone=Llama-3.1-8B-Instruct2026.0557.8276.32
RPSFTBase Model=Llama-3.1-8B-Instruct, RL Stage=DAPO, k=42026.0557.8276.32
SFTBackbone=Llama-3.1-8B-Instruct2026.0556.2971.93
SFTBase Model=Llama-3.1-8B-Instruct, RL Stage=DAPO, k=42026.0556.2971.93
RPSFTBackbone=Qwen2.5-3B-Instruct2026.0552.371.49
RPSFTBase Model=Qwen2.5-3B-Instruct, RL Stage=DAPO, k=42026.0552.371.49
IWBackbone=Qwen2.5-3B-Instruct2026.0550.6969.15
IWBase Model=Qwen2.5-3B-Instruct, RL Stage=DAPO, k=42026.0550.6969.15
SFTBackbone=Qwen2.5-3B-Instruct2026.0550.5571.35
SFTBase Model=Qwen2.5-3B-Instruct, RL Stage=DAPO, k=42026.0550.5571.35
IWBackbone=Llama-3.1-8B-Instruct2026.0550.1866.37
IWBase Model=Llama-3.1-8B-Instruct, RL Stage=DAPO, k=42026.0550.1866.37