bloomz-7b1-p3/evaluation_bloomz-7b1-p3/evaluation_l1/merged.csv

1	dataset	prompt	metric	value
2	anli_dev_r1	GPT-3 style	accuracy	0.351
3	anli_dev_r1	MNLI crowdsource	accuracy	0.334
4	anli_dev_r1	can we infer	accuracy	0.351
5	anli_dev_r1	guaranteed/possible/impossible	accuracy	0.288
6	anli_dev_r1	justified in saying	accuracy	0.345
7	anli_dev_r1	median	accuracy	0.345
8	anli_dev_r2	GPT-3 style	accuracy	0.339
9	anli_dev_r2	MNLI crowdsource	accuracy	0.335
10	anli_dev_r2	can we infer	accuracy	0.354
11	anli_dev_r2	guaranteed/possible/impossible	accuracy	0.297
12	anli_dev_r2	justified in saying	accuracy	0.345
13	anli_dev_r2	median	accuracy	0.339
14	anli_dev_r3	GPT-3 style	accuracy	0.37583333333333335
15	anli_dev_r3	MNLI crowdsource	accuracy	0.3408333333333333
16	anli_dev_r3	can we infer	accuracy	0.36333333333333334
17	anli_dev_r3	guaranteed/possible/impossible	accuracy	0.31083333333333335
18	anli_dev_r3	justified in saying	accuracy	0.34
19	anli_dev_r3	median	accuracy	0.3408333333333333
20	story_cloze_2016	Answer Given options	accuracy	0.8305718866916088
21	story_cloze_2016	Choose Story Ending	accuracy	0.8706574024585783
22	story_cloze_2016	Generate Ending	accuracy	0.7183324425440941
23	story_cloze_2016	Novel Correct Ending	accuracy	0.848743987172635
24	story_cloze_2016	Story Continuation and Options	accuracy	0.8466060929983966
25	story_cloze_2016	median	accuracy	0.8466060929983966
26	super_glue_cb	GPT-3 style	accuracy	0.625
27	super_glue_cb	MNLI crowdsource	accuracy	0.08928571428571429
28	super_glue_cb	can we infer	accuracy	0.5892857142857143
29	super_glue_cb	guaranteed/possible/impossible	accuracy	0.5
30	super_glue_cb	justified in saying	accuracy	0.5357142857142857
31	super_glue_cb	median	accuracy	0.5357142857142857
32	super_glue_copa	C1 or C2? premise, so/because…	accuracy	0.66
33	super_glue_copa	best_option	accuracy	0.67
34	super_glue_copa	cause_effect	accuracy	0.78
35	super_glue_copa	i_am_hesitating	accuracy	0.8
36	super_glue_copa	plausible_alternatives	accuracy	0.81
37	super_glue_copa	median	accuracy	0.78
38	super_glue_rte	GPT-3 style	accuracy	0.7870036101083032
39	super_glue_rte	MNLI crowdsource	accuracy	0.7220216606498195
40	super_glue_rte	does it follow that	accuracy	0.6678700361010831
41	super_glue_rte	guaranteed true	accuracy	0.6714801444043321
42	super_glue_rte	should assume	accuracy	0.6678700361010831
43	super_glue_rte	median	accuracy	0.6714801444043321
44	winogrande_winogrande_xl	Replace	accuracy	0.5406471981057617
45	winogrande_winogrande_xl	True or False	accuracy	0.5074980268350434
46	winogrande_winogrande_xl	does underscore refer to	accuracy	0.5177584846093133
47	winogrande_winogrande_xl	stand for	accuracy	0.510655090765588
48	winogrande_winogrande_xl	underscore refer to	accuracy	0.5256511444356748
49	winogrande_winogrande_xl	median	accuracy	0.5177584846093133
50	xcopa_id	C1 or C2? premise, so/because…	accuracy	0.47
51	xcopa_id	best_option	accuracy	0.51
52	xcopa_id	cause_effect	accuracy	0.65
53	xcopa_id	i_am_hesitating	accuracy	0.66
54	xcopa_id	plausible_alternatives	accuracy	0.67
55	xcopa_id	median	accuracy	0.65
56	xcopa_sw	C1 or C2? premise, so/because…	accuracy	0.58
57	xcopa_sw	best_option	accuracy	0.57
58	xcopa_sw	cause_effect	accuracy	0.46
59	xcopa_sw	i_am_hesitating	accuracy	0.48
60	xcopa_sw	plausible_alternatives	accuracy	0.45
61	xcopa_sw	median	accuracy	0.48
62	xcopa_ta	C1 or C2? premise, so/because…	accuracy	0.57
63	xcopa_ta	best_option	accuracy	0.67
64	xcopa_ta	cause_effect	accuracy	0.71
65	xcopa_ta	i_am_hesitating	accuracy	0.71
66	xcopa_ta	plausible_alternatives	accuracy	0.69
67	xcopa_ta	median	accuracy	0.69
68	xcopa_vi	C1 or C2? premise, so/because…	accuracy	0.55
69	xcopa_vi	best_option	accuracy	0.61
70	xcopa_vi	cause_effect	accuracy	0.67
71	xcopa_vi	i_am_hesitating	accuracy	0.66
72	xcopa_vi	plausible_alternatives	accuracy	0.65
73	xcopa_vi	median	accuracy	0.65
74	xcopa_zh	C1 or C2? premise, so/because…	accuracy	0.62
75	xcopa_zh	best_option	accuracy	0.61
76	xcopa_zh	cause_effect	accuracy	0.77
77	xcopa_zh	i_am_hesitating	accuracy	0.72
78	xcopa_zh	plausible_alternatives	accuracy	0.74
79	xcopa_zh	median	accuracy	0.72
80	xnli_ar	GPT-3 style	accuracy	0.5040160642570282
81	xnli_ar	MNLI crowdsource	accuracy	0.39879518072289155
82	xnli_ar	can we infer	accuracy	0.506425702811245
83	xnli_ar	guaranteed/possible/impossible	accuracy	0.4799196787148594
84	xnli_ar	justified in saying	accuracy	0.41526104417670684
85	xnli_ar	median	accuracy	0.4799196787148594
86	xnli_en	GPT-3 style	accuracy	0.5590361445783133
87	xnli_en	MNLI crowdsource	accuracy	0.342570281124498
88	xnli_en	can we infer	accuracy	0.5449799196787148
89	xnli_en	guaranteed/possible/impossible	accuracy	0.41164658634538154
90	xnli_en	justified in saying	accuracy	0.4634538152610442
91	xnli_en	median	accuracy	0.4634538152610442
92	xnli_es	GPT-3 style	accuracy	0.5373493975903615
93	xnli_es	MNLI crowdsource	accuracy	0.40441767068273093
94	xnli_es	can we infer	accuracy	0.5277108433734939
95	xnli_es	guaranteed/possible/impossible	accuracy	0.44216867469879517
96	xnli_es	justified in saying	accuracy	0.4534136546184739
97	xnli_es	median	accuracy	0.4534136546184739
98	xnli_fr	GPT-3 style	accuracy	0.5248995983935743
99	xnli_fr	MNLI crowdsource	accuracy	0.3895582329317269
100	xnli_fr	can we infer	accuracy	0.5337349397590362
101	xnli_fr	guaranteed/possible/impossible	accuracy	0.42971887550200805
102	xnli_fr	justified in saying	accuracy	0.4738955823293173
103	xnli_fr	median	accuracy	0.4738955823293173
104	xnli_hi	GPT-3 style	accuracy	0.4983935742971888
105	xnli_hi	MNLI crowdsource	accuracy	0.38714859437751004
106	xnli_hi	can we infer	accuracy	0.45542168674698796
107	xnli_hi	guaranteed/possible/impossible	accuracy	0.41405622489959837
108	xnli_hi	justified in saying	accuracy	0.38795180722891565
109	xnli_hi	median	accuracy	0.41405622489959837
110	xnli_sw	GPT-3 style	accuracy	0.43493975903614457
111	xnli_sw	MNLI crowdsource	accuracy	0.363855421686747
112	xnli_sw	can we infer	accuracy	0.42891566265060244
113	xnli_sw	guaranteed/possible/impossible	accuracy	0.3457831325301205
114	xnli_sw	justified in saying	accuracy	0.3650602409638554
115	xnli_sw	median	accuracy	0.3650602409638554
116	xnli_ur	GPT-3 style	accuracy	0.43493975903614457
117	xnli_ur	MNLI crowdsource	accuracy	0.3895582329317269
118	xnli_ur	can we infer	accuracy	0.45180722891566266
119	xnli_ur	guaranteed/possible/impossible	accuracy	0.40120481927710844
120	xnli_ur	justified in saying	accuracy	0.37630522088353413
121	xnli_ur	median	accuracy	0.40120481927710844
122	xnli_vi	GPT-3 style	accuracy	0.5196787148594377
123	xnli_vi	MNLI crowdsource	accuracy	0.38112449799196785
124	xnli_vi	can we infer	accuracy	0.5080321285140562
125	xnli_vi	guaranteed/possible/impossible	accuracy	0.38393574297188754
126	xnli_vi	justified in saying	accuracy	0.43614457831325304
127	xnli_vi	median	accuracy	0.43614457831325304
128	xnli_zh	GPT-3 style	accuracy	0.5052208835341365
129	xnli_zh	MNLI crowdsource	accuracy	0.4
130	xnli_zh	can we infer	accuracy	0.5228915662650603
131	xnli_zh	guaranteed/possible/impossible	accuracy	0.4738955823293173
132	xnli_zh	justified in saying	accuracy	0.45863453815261046
133	xnli_zh	median	accuracy	0.4738955823293173
134	xstory_cloze_ar	Answer Given options	accuracy	0.7518199867637326
135	xstory_cloze_ar	Choose Story Ending	accuracy	0.7749834546657842
136	xstory_cloze_ar	Generate Ending	accuracy	0.586366644606221
137	xstory_cloze_ar	Novel Correct Ending	accuracy	0.7518199867637326
138	xstory_cloze_ar	Story Continuation and Options	accuracy	0.7438782263401721
139	xstory_cloze_ar	median	accuracy	0.7518199867637326
140	xstory_cloze_es	Answer Given options	accuracy	0.7835870284579749
141	xstory_cloze_es	Choose Story Ending	accuracy	0.8292521508934481
142	xstory_cloze_es	Generate Ending	accuracy	0.6399735274652548
143	xstory_cloze_es	Novel Correct Ending	accuracy	0.7935142289874255
144	xstory_cloze_es	Story Continuation and Options	accuracy	0.7888815354070152
145	xstory_cloze_es	median	accuracy	0.7888815354070152
146	xstory_cloze_eu	Answer Given options	accuracy	0.7041694242223693
147	xstory_cloze_eu	Choose Story Ending	accuracy	0.6823295830575777
148	xstory_cloze_eu	Generate Ending	accuracy	0.5625413633355394
149	xstory_cloze_eu	Novel Correct Ending	accuracy	0.6671078755790867
150	xstory_cloze_eu	Story Continuation and Options	accuracy	0.671740569159497
151	xstory_cloze_eu	median	accuracy	0.671740569159497
152	xstory_cloze_hi	Answer Given options	accuracy	0.6915949702183984
153	xstory_cloze_hi	Choose Story Ending	accuracy	0.7220383851753805
154	xstory_cloze_hi	Generate Ending	accuracy	0.5883520847121112
155	xstory_cloze_hi	Novel Correct Ending	accuracy	0.6743878226340172
156	xstory_cloze_hi	Story Continuation and Options	accuracy	0.6816677696889477
157	xstory_cloze_hi	median	accuracy	0.6816677696889477
158	xstory_cloze_id	Answer Given options	accuracy	0.7445400397088021
159	xstory_cloze_id	Choose Story Ending	accuracy	0.771012574454004
160	xstory_cloze_id	Generate Ending	accuracy	0.6029119788219722
161	xstory_cloze_id	Novel Correct Ending	accuracy	0.7485109199205824
162	xstory_cloze_id	Story Continuation and Options	accuracy	0.7438782263401721
163	xstory_cloze_id	median	accuracy	0.7445400397088021
164	xstory_cloze_zh	Answer Given options	accuracy	0.7610853739245532
165	xstory_cloze_zh	Choose Story Ending	accuracy	0.7961614824619457
166	xstory_cloze_zh	Generate Ending	accuracy	0.6214427531436135
167	xstory_cloze_zh	Novel Correct Ending	accuracy	0.7696889477167439
168	xstory_cloze_zh	Story Continuation and Options	accuracy	0.7670416942422237
169	xstory_cloze_zh	median	accuracy	0.7670416942422237
170	xwinograd_en	Replace	accuracy	0.5225806451612903
171	xwinograd_en	True or False	accuracy	0.48946236559139783
172	xwinograd_en	does underscore refer to	accuracy	0.5281720430107527
173	xwinograd_en	stand for	accuracy	0.5062365591397849
174	xwinograd_en	underscore refer to	accuracy	0.5372043010752688
175	xwinograd_en	median	accuracy	0.5225806451612903
176	xwinograd_fr	Replace	accuracy	0.5060240963855421
177	xwinograd_fr	True or False	accuracy	0.5421686746987951
178	xwinograd_fr	does underscore refer to	accuracy	0.5542168674698795
179	xwinograd_fr	stand for	accuracy	0.4819277108433735
180	xwinograd_fr	underscore refer to	accuracy	0.5301204819277109
181	xwinograd_fr	median	accuracy	0.5301204819277109
182	xwinograd_pt	Replace	accuracy	0.5133079847908745
183	xwinograd_pt	True or False	accuracy	0.4714828897338403
184	xwinograd_pt	does underscore refer to	accuracy	0.5209125475285171
185	xwinograd_pt	stand for	accuracy	0.5019011406844106
186	xwinograd_pt	underscore refer to	accuracy	0.5399239543726235
187	xwinograd_pt	median	accuracy	0.5133079847908745
188	xwinograd_zh	Replace	accuracy	0.5257936507936508
189	xwinograd_zh	True or False	accuracy	0.5297619047619048
190	xwinograd_zh	does underscore refer to	accuracy	0.5218253968253969
191	xwinograd_zh	stand for	accuracy	0.4444444444444444
192	xwinograd_zh	underscore refer to	accuracy	0.5198412698412699
193	xwinograd_zh	median	accuracy	0.5218253968253969
194	multiple	average	multiple	0.5631550819200618