Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Factual Consistency Detection on TofuEval-MeetingBank Response level leakage-clean protocol (test)
Loading...
65.9
AUC
GASP+base (trained)
51.028
54.889
58.75
62.611
Jul 5, 2026
AUC
Updated 18d ago
Evaluation Results
Method
Method
Links
AUC
GASP+base (trained)
Model=Qwen-0.5B
2026.07
65.9
GASP+base (trained)
Model=SmolLM2
2026.07
65.6
GASP-trained
Model=SmolLM2
2026.07
64.1
GASP+base (trained)
Model=Qwen-1.5B
2026.07
63.3
GASP-trained
Model=Qwen-0.5B
2026.07
63.2
GASP-trained
Model=Qwen-1.5B
2026.07
62.7
GASP-threshold (default)
Model=Qwen-0.5B
2026.07
62.5
GASP-threshold (default)
Model=Qwen-1.5B
2026.07
62.3
GASP-threshold (default)
Model=SmolLM2
2026.07
61.4
Perplexity
Model=Qwen-1.5B
2026.07
54.8
Perplexity
Model=Qwen-0.5B
2026.07
52.1
Perplexity
Model=SmolLM2
2026.07
51.6
Feedback
Search any
task
Search any
task