DailyDialog
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
DailyDialog (test)
0.603Micro-F1
53
DailyDialog
0.617Micro F1 (- neutral)
27
DailyDialog
9.12Distinct-1
26
DailyDialog
0.54BERT Score
18
DailyDialog
84.29BERTScore Precision (avg)
16
DailyDialog reformatted multiple-choice (test)
90.35Accuracy
16
DailyDialog Multi-reference
38.46BLEU-1
16
DailyDialog (test)
35.4BLEU-2
16
DailyDialog (test)
0.631F1 Score
16
DailyDialog
59.33Macro F1
15
DailyDialog-CG (test)
70.66Emotion Accuracy (E-ACC)
12
DailyDialog (eval)
0.579Spearman Correlation
10
DailyDialog
14.99R-1
10
DailyDialog (test)
89.58Precision
9
DailyDialog
80.97Human Logic Alignment (T=0.5)
9
DailyDialog
0.57Reward
8
DailyDialog against DialoGPT-large
40RSR
8
DailyDialog against BB-3B
40.2RSR
8
Dailydialog (test)
81.1USR MLM
8
DailyDialog multi-reference (test)
80.25Accuracy
7
DailyDialog (DD) (Fold 1)
69.2F1 (Positive)
7
DailyDialog out-of-domain
12.01ROUGE-L
6
DailyDialog (test)
12.53Accuracy
6
Dailydialog
34.7Attribute Relevancy
6
DailyDialog
78.5Pairwise Diversity
6