List your company

ResearchFormerly WizWand

OpenDialKG

Benchmarks

TaskDatasetSOTA resultTrendResults

Knowledge-intensive Dialogue

OpendialKG

88.84Factual Accuracy

11

Knowledge-grounded dialogue generation

OpenDialKG (test)

20.77BLEU-4

10

Hallucination detection

OpenDialKG Eval (test)

76.2Macro F1

7

OpenDialKG

2.916Dist-2

7

OpenDialKG

28.95Recall@1

7

RDF-to-text generation

OpenDialKG (test)

98.5Grammaticality

6

Knowledge-grounded Dialogue Generation

OpenDialKG

81.67Faithfulness

4

Knowledge-Grounded Dialogue Generation (Fluency)

OpenDialKG

37.33Win Rate

4