List your company

Hallucination Detection on RAGTruth Span-level, leakage-clean protocol

0.702AUCGASP+base (trained)
0.515840.564170.61250.66083Jul 5, 2026

Evaluation results

MethodLinks
GASP+base (trained)Scorer=SmolLM2-1.7B, Protocol=leakage-clean2026.070.702
GASP-threshold (default)Scorer=SmolLM2-1.7B, Protocol=leakage-clean2026.070.681
GASP+base (trained)Scorer=Qwen2.5-1.5B, Protocol=leakage-clean2026.070.678
GASP+base (trained)Scorer=Qwen2.5-0.5B, Protocol=leakage-clean2026.070.676
GASP-threshold (default)Scorer=Qwen2.5-1.5B, Protocol=leakage-clean2026.070.673
GASP-threshold (default)Scorer=Qwen2.5-0.5B, Protocol=leakage-clean2026.070.672
GASP-trainedScorer=SmolLM2-1.7B, Protocol=leakage-clean2026.070.657
GASP-trainedScorer=Qwen2.5-1.5B, Protocol=leakage-clean2026.070.645
GASP-trainedScorer=Qwen2.5-0.5B, Protocol=leakage-clean2026.070.635
PerplexityScorer=SmolLM2-1.7B, Protocol=leakage-clean2026.070.615
PerplexityScorer=Qwen2.5-1.5B, Protocol=leakage-clean2026.070.565
LengthScorer=Qwen2.5-1.5B, Protocol=leakage-clean2026.070.548
LengthScorer=Qwen2.5-0.5B, Protocol=leakage-clean2026.070.547
PerplexityScorer=Qwen2.5-0.5B, Protocol=leakage-clean2026.070.536
LengthScorer=SmolLM2-1.7B, Protocol=leakage-clean2026.070.523