bloomz-7b1-p3/evaluation_bloomz-7b1-p3/evaluation_xnliht/xnli/merged.csv

1	dataset	prompt	metric	value
2	xnli_ar	GPT-3 style_arht	accuracy	0.3634538152610442
3	xnli_ar	MNLI crowdsource_arht	accuracy	0.3437751004016064
4	xnli_ar	can we infer_arht	accuracy	0.40923694779116465
5	xnli_ar	guaranteed/possible/impossible_arht	accuracy	0.342570281124498
6	xnli_ar	justified in saying_arht	accuracy	0.4321285140562249
7	xnli_ar	median	accuracy	0.3634538152610442
8	xnli_es	GPT-3 style_esht	accuracy	0.4795180722891566
9	xnli_es	MNLI crowdsource_esht	accuracy	0.3333333333333333
10	xnli_es	can we infer_esht	accuracy	0.3333333333333333
11	xnli_es	guaranteed/possible/impossible_esht	accuracy	0.529718875502008
12	xnli_es	justified in saying_esht	accuracy	0.3333333333333333
13	xnli_es	median	accuracy	0.3333333333333333
14	xnli_fr	GPT-3 style_frht	accuracy	0.45863453815261046
15	xnli_fr	MNLI crowdsource_frht	accuracy	0.42730923694779116
16	xnli_fr	can we infer_frht	accuracy	0.40963855421686746
17	xnli_fr	guaranteed/possible/impossible_frht	accuracy	0.37309236947791163
18	xnli_fr	justified in saying_frht	accuracy	0.4710843373493976
19	xnli_fr	median	accuracy	0.42730923694779116
20	xnli_hi	GPT-3 style_hiht	accuracy	0.3542168674698795
21	xnli_hi	MNLI crowdsource_hiht	accuracy	0.4389558232931727
22	xnli_hi	can we infer_hiht	accuracy	0.41566265060240964
23	xnli_hi	guaranteed/possible/impossible_hiht	accuracy	0.4927710843373494
24	xnli_hi	justified in saying_hiht	accuracy	0.4562248995983936
25	xnli_hi	median	accuracy	0.4389558232931727
26	xnli_sw	GPT-3 style_swht	accuracy	0.3389558232931727
27	xnli_sw	MNLI crowdsource_swht	accuracy	0.3257028112449799
28	xnli_sw	can we infer_swht	accuracy	0.3429718875502008
29	xnli_sw	guaranteed/possible/impossible_swht	accuracy	0.3718875502008032
30	xnli_sw	justified in saying_swht	accuracy	0.3409638554216867
31	xnli_sw	median	accuracy	0.3409638554216867
32	xnli_ur	GPT-3 style_urht	accuracy	0.3646586345381526
33	xnli_ur	MNLI crowdsource_urht	accuracy	0.3538152610441767
34	xnli_ur	can we infer_urht	accuracy	0.3610441767068273
35	xnli_ur	guaranteed/possible/impossible_urht	accuracy	0.37670682730923694
36	xnli_ur	justified in saying_urht	accuracy	0.3377510040160643
37	xnli_ur	median	accuracy	0.3610441767068273
38	xnli_vi	GPT-3 style_viht	accuracy	0.3357429718875502
39	xnli_vi	MNLI crowdsource_viht	accuracy	0.3477911646586345
40	xnli_vi	can we infer_viht	accuracy	0.3333333333333333
41	xnli_vi	guaranteed/possible/impossible_viht	accuracy	0.39759036144578314
42	xnli_vi	justified in saying_viht	accuracy	0.3333333333333333
43	xnli_vi	median	accuracy	0.3357429718875502
44	xnli_zh	GPT-3 style_zhht	accuracy	0.348995983935743
45	xnli_zh	MNLI crowdsource_zhht	accuracy	0.45100401606425705
46	xnli_zh	can we infer_zhht	accuracy	0.3931726907630522
47	xnli_zh	guaranteed/possible/impossible_zhht	accuracy	0.39879518072289155
48	xnli_zh	justified in saying_zhht	accuracy	0.3827309236947791
49	xnli_zh	median	accuracy	0.3931726907630522
50	multiple	average	multiple	0.3742469879518072