| open jev / customer control v1 |
Choice |
69.6 |
82.6 |
71.5 |
|
来源 ↗ |
| laya / ag news |
Choice |
68.7 |
94.0 |
77.7 |
|
来源 ↗ |
| open jev / mailroom control v1 |
Choice |
66.7 |
100.0 |
89.6 |
|
来源 ↗ |
| dbpedia |
Choice |
65.2 |
98.9 |
88.9 |
|
来源 ↗ |
| open jev / ir control v1 |
Noul |
64.3 |
100.0 |
60.3 |
|
来源 ↗ |
| open jev / mailroom control v1 |
Noul |
60.0 |
100.0 |
73.2 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanohotpotqa |
Noul |
58.5 |
78.9 |
36.1 |
|
来源 ↗ |
| scienceqa |
Choice |
56.8 |
99.5 |
86.7 |
|
来源 ↗ |
| generalization vision easy |
Choice |
56.7 |
98.6 |
90.7 |
|
来源 ↗ |
| nights |
Choice |
56.4 |
92.0 |
75.3 |
|
来源 ↗ |
| followir robust04 |
Noul |
54.9 |
90.9 |
53.7 |
|
来源 ↗ |
| s1mb generalization contextual choice |
Choice |
52.4 |
100.0 |
88.7 |
|
来源 ↗ |
| sugarcrepe |
Choice |
51.6 |
95.2 |
80.9 |
|
来源 ↗ |
| open jev / workflow controls v1 / customer service |
Noul |
51.1 |
100.0 |
53.7 |
|
来源 ↗ |
| cvbench relation |
Choice |
47.2 |
96.5 |
79.7 |
|
来源 ↗ |
| open jev / customer control v1 |
Noul |
46.7 |
100.0 |
68.3 |
|
来源 ↗ |
| hans |
Noul |
44.0 |
100.0 |
53.1 |
|
来源 ↗ |
| open jev / silent failure control v1 |
Noul |
44.0 |
100.0 |
49.8 |
|
来源 ↗ |
| sdoh nli |
Noul |
42.0 |
100.0 |
70.0 |
|
来源 ↗ |
| go emotions |
Noul |
40.8 |
63.4 |
42.1 |
|
来源 ↗ |
| s1mb generalization diverse choice |
Choice |
40.5 |
100.0 |
78.1 |
|
来源 ↗ |
| open jev / workflow controls v1 / security incidents |
Noul |
40.4 |
100.0 |
50.0 |
|
来源 ↗ |
| generalization vision easy |
Noul |
39.7 |
100.0 |
87.4 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanonq |
Noul |
36.3 |
72.7 |
48.0 |
|
来源 ↗ |
| generalization japanese diverse |
Noul |
34.7 |
100.0 |
91.4 |
|
来源 ↗ |
| canttalk |
Noul |
34.0 |
92.0 |
50.5 |
|
来源 ↗ |
| s1mb generalization contextual noul |
Noul |
34.0 |
100.0 |
72.1 |
|
来源 ↗ |
| few nerd |
Choice |
31.9 |
87.8 |
52.9 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanodbpedia |
Noul |
31.5 |
53.2 |
33.8 |
|
来源 ↗ |
| open jev / email selection control v1 |
Noul |
30.8 |
100.0 |
56.9 |
|
来源 ↗ |
| followir news21 |
Noul |
30.0 |
58.0 |
18.4 |
|
来源 ↗ |
| open jev / phone extraction control v1 |
Noul |
30.0 |
100.0 |
63.9 |
|
来源 ↗ |
| s1mb generalization diverse noul |
Noul |
30.0 |
100.0 |
70.7 |
|
来源 ↗ |
| mvtec ad |
Noul |
29.0 |
76.9 |
51.0 |
|
来源 ↗ |
| ethos |
Noul |
28.0 |
76.0 |
49.5 |
|
来源 ↗ |
| janli |
Choice |
27.4 |
88.9 |
41.7 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanomsmarco |
Noul |
26.1 |
64.1 |
45.3 |
|
来源 ↗ |
| impli |
Noul |
26.0 |
96.0 |
57.1 |
|
来源 ↗ |
| open jev / reasoning control v1 |
Noul |
25.4 |
100.0 |
34.3 |
|
来源 ↗ |
| ethics |
Noul |
24.9 |
80.6 |
33.4 |
|
来源 ↗ |
| visa |
Noul |
24.8 |
73.4 |
29.3 |
|
来源 ↗ |
| koniq |
Score |
24.6 |
51.8 |
8.2 |
|
来源 ↗ |
| open jev / context retention control v1 |
Noul |
24.0 |
100.0 |
35.2 |
|
来源 ↗ |
| followir core17 |
Noul |
23.0 |
46.6 |
25.2 |
|
来源 ↗ |
| arc |
Choice |
22.1 |
98.5 |
73.8 |
|
来源 ↗ |
| babi nli |
Noul |
22.0 |
96.0 |
48.5 |
|
来源 ↗ |
| synthetic relevance / nanocoir |
Noul |
21.2 |
73.6 |
42.2 |
|
来源 ↗ |
| civil comments |
Noul |
19.7 |
52.8 |
25.2 |
|
来源 ↗ |
| open jev / workflow controls v1 / agent trace observability |
Noul |
18.4 |
100.0 |
39.4 |
|
来源 ↗ |
| ruletaker |
Noul |
18.0 |
100.0 |
39.3 |
|
来源 ↗ |
| scitail |
Noul |
18.0 |
98.0 |
63.4 |
|
来源 ↗ |
| open jev / sponsor segment control v1 |
Choice |
17.9 |
100.0 |
66.4 |
|
来源 ↗ |
| pope |
Noul |
16.7 |
80.0 |
73.0 |
|
来源 ↗ |
| mtop |
Choice |
16.6 |
93.2 |
67.4 |
|
来源 ↗ |
| plane |
Noul |
16.0 |
100.0 |
12.6 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanonfcorpus |
Noul |
15.5 |
49.7 |
29.6 |
|
来源 ↗ |
| openbookqa |
Choice |
15.3 |
98.6 |
64.4 |
|
来源 ↗ |
| rvlcdip |
Choice |
14.9 |
77.6 |
61.4 |
|
来源 ↗ |
| jcommonsenseqa |
Choice |
14.2 |
97.9 |
87.6 |
|
来源 ↗ |
| laya / typed decisions |
Choice |
13.8 |
35.7 |
23.3 |
|
来源 ↗ |
| open jev / amount extraction control v1 |
Choice |
13.7 |
100.0 |
63.2 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanofiqa2018 |
Noul |
12.7 |
58.0 |
35.3 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanoscidocs |
Noul |
12.6 |
48.5 |
21.5 |
|
来源 ↗ |
| generalization japanese diverse |
Choice |
12.5 |
100.0 |
94.2 |
|
来源 ↗ |
| hatecheck |
Noul |
12.0 |
100.0 |
57.5 |
|
来源 ↗ |
| jamp |
Choice |
11.5 |
81.5 |
50.0 |
|
来源 ↗ |
| massive |
Choice |
10.6 |
97.8 |
73.4 |
|
来源 ↗ |
| sgd |
Choice |
10.2 |
96.6 |
28.6 |
|
来源 ↗ |
| jnli |
Choice |
10.0 |
87.5 |
68.4 |
|
来源 ↗ |
| winowhy |
Noul |
10.0 |
66.0 |
25.9 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanotouche2020 |
Noul |
10.0 |
48.2 |
32.3 |
|
来源 ↗ |
| snli |
Choice |
9.7 |
98.4 |
59.9 |
|
来源 ↗ |
| open jev / amount extraction control v1 |
Noul |
9.4 |
100.0 |
44.2 |
|
来源 ↗ |
| generalization japanese knowledge hard |
Choice |
8.9 |
99.6 |
88.0 |
|
来源 ↗ |
| open jev / workflow controls v1 / invoice processing |
Noul |
8.6 |
100.0 |
47.0 |
|
来源 ↗ |
| kuci |
Choice |
8.4 |
85.8 |
64.3 |
|
来源 ↗ |
| bbq |
Choice |
8.3 |
98.3 |
64.5 |
|
来源 ↗ |
| jsts |
Score |
7.6 |
69.9 |
55.3 |
|
来源 ↗ |
| nlvr2 |
Noul |
7.3 |
88.9 |
59.2 |
|
来源 ↗ |
| temporal nli |
Choice |
6.5 |
100.0 |
28.0 |
|
来源 ↗ |
| hwu64 |
Choice |
6.4 |
95.7 |
67.3 |
|
来源 ↗ |
| banking77 |
Choice |
6.3 |
93.7 |
64.2 |
|
来源 ↗ |
| clinc |
Choice |
6.2 |
99.3 |
70.5 |
|
来源 ↗ |
| cladder |
Noul |
6.0 |
94.0 |
28.5 |
|
来源 ↗ |
| logical entailment |
Noul |
6.0 |
84.0 |
18.7 |
|
来源 ↗ |
| jmmlu japan |
Choice |
5.8 |
94.2 |
68.4 |
|
来源 ↗ |
| generalization japanese math formal hard |
Noul |
5.3 |
34.7 |
10.6 |
|
来源 ↗ |
| open jev / snake v1 |
Noul |
5.1 |
100.0 |
19.1 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanoarguana |
Noul |
4.1 |
53.5 |
22.0 |
|
来源 ↗ |
| qasper |
Noul |
4.0 |
88.0 |
51.4 |
|
来源 ↗ |
| generalization vision hard |
Noul |
3.7 |
84.9 |
47.9 |
|
来源 ↗ |
| gpqa ja main |
Choice |
3.6 |
54.7 |
20.9 |
|
来源 ↗ |
| jmteb retrieval |
Choice |
3.4 |
98.7 |
90.0 |
|
来源 ↗ |
| jamc qa |
Choice |
2.7 |
65.3 |
23.0 |
|
来源 ↗ |
| mintaka ja boolean |
Noul |
2.7 |
68.0 |
37.5 |
|
来源 ↗ |
| generalization japanese math formal hard |
Choice |
2.2 |
48.4 |
21.0 |
|
来源 ↗ |
| creak |
Noul |
2.0 |
90.0 |
48.2 |
|
来源 ↗ |
| s1mb generalization diverse score |
Score |
1.9 |
97.6 |
52.7 |
|
来源 ↗ |
| s1mb generalization contextual score |
Score |
1.3 |
99.3 |
52.6 |
|
来源 ↗ |
| stepgame |
Choice |
1.2 |
82.1 |
24.8 |
|
来源 ↗ |
| open jev / reasoning control v1 |
Score |
1.0 |
96.3 |
38.3 |
|
来源 ↗ |
| mmmlu |
Choice |
0.9 |
78.2 |
52.8 |
|
来源 ↗ |
| commonsensemoralja |
Noul |
0.7 |
89.3 |
67.8 |
|
来源 ↗ |
| cvbench depth |
Choice |
0.7 |
93.3 |
77.6 |
|
来源 ↗ |
| aegis2 |
Noul |
0.0 |
68.3 |
39.6 |
|
来源 ↗ |
| aqua rat |
Choice |
0.0 |
83.1 |
19.7 |
|
来源 ↗ |
| arct |
Choice |
0.0 |
89.4 |
49.3 |
|
来源 ↗ |
| argument quality |
Choice |
0.0 |
95.9 |
53.3 |
|
来源 ↗ |
| argument quality |
Noul |
0.0 |
30.2 |
7.8 |
|
来源 ↗ |
| boardgameqa |
Choice |
0.0 |
90.6 |
24.4 |
|
来源 ↗ |
| contract nli |
Choice |
0.0 |
81.6 |
43.0 |
|
来源 ↗ |
| corr2cause |
Choice |
0.0 |
91.7 |
9.3 |
|
来源 ↗ |
| crows pairs |
Choice |
0.0 |
77.1 |
37.9 |
|
来源 ↗ |
| cvbench distance |
Choice |
0.0 |
79.9 |
59.3 |
|
来源 ↗ |
| defeasible nli |
Choice |
0.0 |
89.1 |
43.1 |
|
来源 ↗ |
| esci |
Choice |
0.0 |
49.3 |
17.5 |
|
来源 ↗ |
| ethics |
Choice |
0.0 |
100.0 |
28.5 |
|
来源 ↗ |
| fol nli |
Choice |
0.0 |
75.4 |
17.3 |
|
来源 ↗ |
| generalization japanese knowledge hard |
Noul |
0.0 |
87.3 |
50.6 |
|
来源 ↗ |
| generalization japanese knowledge hard |
Score |
0.0 |
84.6 |
45.3 |
|
来源 ↗ |
| generalization japanese math formal hard |
Score |
0.0 |
6.9 |
0.9 |
|
来源 ↗ |
| generalization vision easy |
Score |
0.0 |
79.7 |
50.0 |
|
来源 ↗ |
| generalization vision hard |
Choice |
0.0 |
84.1 |
48.5 |
|
来源 ↗ |
| generalization vision hard |
Score |
0.0 |
49.2 |
19.9 |
|
来源 ↗ |
| gretel pii |
Choice |
0.0 |
94.8 |
51.1 |
|
来源 ↗ |
| gsm8k |
Choice |
0.0 |
95.7 |
38.0 |
|
来源 ↗ |
| hh rlhf |
Choice |
0.0 |
33.3 |
4.5 |
|
来源 ↗ |
| jcola |
Noul |
0.0 |
57.2 |
21.6 |
|
来源 ↗ |
| jculture mcq |
Noul |
0.0 |
44.3 |
12.8 |
|
来源 ↗ |
| jsem |
Choice |
0.0 |
69.2 |
45.9 |
|
来源 ↗ |
| jsick sts |
Score |
0.0 |
61.5 |
40.1 |
|
来源 ↗ |
| laya / enron spam |
Noul |
0.0 |
100.0 |
64.9 |
|
来源 ↗ |
| laya / typed decisions |
Noul |
0.0 |
49.2 |
29.9 |
|
来源 ↗ |
| laya / typed decisions |
Score |
0.0 |
76.0 |
35.4 |
|
来源 ↗ |
| lexcomp |
Noul |
0.0 |
60.0 |
27.7 |
|
来源 ↗ |
| lonli |
Choice |
0.0 |
98.4 |
52.8 |
|
来源 ↗ |
| miqa |
Choice |
0.0 |
97.7 |
71.1 |
|
来源 ↗ |
| mmlu prox ja |
Choice |
0.0 |
69.5 |
39.5 |
|
来源 ↗ |
| nlsat |
Noul |
0.0 |
22.0 |
3.2 |
|
来源 ↗ |
| open jev / browser control v1 |
Choice |
0.0 |
100.0 |
76.5 |
|
来源 ↗ |
| open jev / citation control v1 |
Choice |
0.0 |
100.0 |
55.9 |
|
来源 ↗ |
| open jev / customer control v1 |
Score |
0.0 |
99.3 |
44.4 |
|
来源 ↗ |
| open jev / drone control v1 |
Choice |
0.0 |
100.0 |
13.4 |
|
来源 ↗ |
| open jev / drone control v1 |
Noul |
0.0 |
100.0 |
63.8 |
|
来源 ↗ |
| open jev / drone control v1 |
Score |
0.0 |
100.0 |
7.2 |
|
来源 ↗ |
| open jev / email selection control v1 |
Choice |
0.0 |
100.0 |
38.4 |
|
来源 ↗ |
| open jev / entity alignment control v1 |
Noul |
0.0 |
100.0 |
50.6 |
|
来源 ↗ |
| open jev / entity alignment control v1 |
Score |
0.0 |
100.0 |
18.4 |
|
来源 ↗ |
| open jev / ir control v1 |
Choice |
0.0 |
88.0 |
55.5 |
|
来源 ↗ |
| open jev / ir control v1 |
Score |
0.0 |
100.0 |
23.0 |
|
来源 ↗ |
| open jev / painting geometry v1 |
Choice |
0.0 |
100.0 |
60.5 |
|
来源 ↗ |
| open jev / painting geometry v1 |
Noul |
0.0 |
100.0 |
42.0 |
|
来源 ↗ |
| open jev / painting geometry v1 |
Score |
0.0 |
100.0 |
29.7 |
|
来源 ↗ |
| open jev / phone extraction control v1 |
Choice |
0.0 |
100.0 |
44.4 |
|
来源 ↗ |
| open jev / reasoning control v1 |
Choice |
0.0 |
100.0 |
41.7 |
|
来源 ↗ |
| open jev / snake v1 |
Choice |
0.0 |
83.5 |
18.3 |
|
来源 ↗ |
| open jev / tic tac toe v1 |
Choice |
0.0 |
54.5 |
5.1 |
|
来源 ↗ |
| open jev / vizdoom basic v1 |
Noul |
0.0 |
100.0 |
53.2 |
|
来源 ↗ |
| open jev / vizdoom basic v1 |
Score |
0.0 |
100.0 |
42.3 |
|
来源 ↗ |
| patent similarity |
Score |
0.0 |
53.0 |
24.0 |
|
来源 ↗ |
| paws |
Noul |
0.0 |
92.0 |
52.5 |
|
来源 ↗ |
| poem sentiment |
Choice |
0.0 |
68.6 |
9.3 |
|
来源 ↗ |
| quartz |
Choice |
0.0 |
90.5 |
59.4 |
|
来源 ↗ |
| robust lr |
Choice |
0.0 |
98.3 |
29.3 |
|
来源 ↗ |
| scicite |
Choice |
0.0 |
74.2 |
49.4 |
|
来源 ↗ |
| scone |
Choice |
0.0 |
98.0 |
48.5 |
|
来源 ↗ |
| sobaco culture |
Choice |
0.0 |
80.5 |
30.6 |
|
来源 ↗ |
| spartqa |
Choice |
0.0 |
93.5 |
14.0 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanoarguana |
Score |
0.0 |
43.2 |
10.3 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanodbpedia |
Score |
0.0 |
50.3 |
23.4 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanofiqa2018 |
Score |
0.0 |
49.0 |
21.0 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanohotpotqa |
Score |
0.0 |
71.8 |
20.9 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanomsmarco |
Score |
0.0 |
57.3 |
25.0 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanonfcorpus |
Score |
0.0 |
40.9 |
15.8 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanonq |
Score |
0.0 |
54.3 |
13.3 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanoscidocs |
Score |
0.0 |
36.7 |
10.7 |
|
来源 ↗ |
| synthetic relevance / nanobeir / nanotouche2020 |
Score |
0.0 |
49.4 |
25.5 |
|
来源 ↗ |
| synthetic relevance / nanocoir |
Score |
0.0 |
43.8 |
6.8 |
|
来源 ↗ |
| tallyqa |
Score |
0.0 |
72.0 |
48.0 |
|
来源 ↗ |
| tracie |
Noul |
0.0 |
68.0 |
14.4 |
|
来源 ↗ |
| ud ewt |
Choice |
0.0 |
96.0 |
19.6 |
|
来源 ↗ |
| vsr |
Noul |
0.0 |
77.6 |
48.0 |
|
来源 ↗ |
| winobias |
Choice |
0.0 |
95.7 |
42.7 |
|
来源 ↗ |
| wiqa |
Choice |
0.0 |
68.3 |
24.0 |
|
来源 ↗ |