Explanation Generation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
37.13ROUGE-1
30
Feb 26, 2026
6.28BLEU-4
19
Jun 30, 2026
10.6BLEU-4
17
Apr 21, 2026
45.98BLEU-4
17
Feb 26, 2026
34.46BLEU-4
17
Feb 26, 2026
40.54BLEU-4
17
Feb 26, 2026
86.1ECS
16
Jul 7, 2026
0.1Per-instance Explanation Time (s)
15
Mar 19, 2026
7.39BLEU-4
12
Jun 30, 2026
4.528BLEU-4
12
Jun 30, 2026
3.14Informativeness
11
May 26, 2026
25.5ROU-1
11
Feb 26, 2026
37.62ROUGE-1
10
Feb 26, 2026
93.26B Score
9
Jun 23, 2026
48.9ROUGE-1
9
Feb 26, 2026
42.9ROUGE-1
9
Feb 26, 2026
40.8ROUGE-L
9
Feb 26, 2026
34.5ROUGE-L
9
Feb 26, 2026
48.7ROUGE-1
9
Feb 26, 2026
42.7ROUGE-1
9
Feb 26, 2026
84.7Faithfulness
7
Apr 24, 2026
4.56PPL
7
Apr 21, 2026
4.14PPL
7
Apr 21, 2026
4.3PPL
7
Apr 21, 2026
5.05PPL
7
Apr 21, 2026