Skip to content

Benchmark

HF Hallucinations HaluEval Dialog, EM

Factuality benchmarks measure how often models get facts right. Higher scores are better.

Models measured
2
Scores recorded
2
Sources
1
IQ weight
No weight published

Score distribution

0.649
Mean
0.655
Median
0.006
Std dev
0.643
10th percentile
0.655
90th percentile
0.643
Min

Leaderboard

#ModelScore
1Llama-2-13b-chat-hfclustermap_all/HaluEval Dialog, EM· HF Hallucinations aggregate matrix0.655
2Llama-2-7b-chat-hfclustermap_all/HaluEval Dialog, EM· HF Hallucinations aggregate matrix0.643

Data credits: Hallucinations Leaderboard, Apache-2.0. Publisher-selected archived score matrix; native normalized fractions. (licence)