List your company

Hallucination Detection on HaluEval (HaluEval Δ Metric)

21.3HaluEval DeltaSTM
-0.8524.89910.6516.401May 19, 2026

Evaluation results

MethodLinks
STMModel=Qwen3-4B2026.0521.3
SFTModel=Qwen3-4B2026.0518.8
FLOWModel=Qwen3-4B2026.0518.8
L2 RegModel=Qwen3-4B2026.0518.8
TALRModel=Qwen3-4B2026.0511.2
LoRAModel=Qwen3-4B2026.058.2
FINCHModel=Qwen3-4B2026.056.7
WiSE-FTModel=Qwen3-4B2026.052.5
DFTModel=Qwen3-4B2026.050.6
BaseModel=Qwen3-4B2026.050