ronnengmail commited on
Commit
cbe7577
·
verified ·
1 Parent(s): 4b368cd

Upload exp_ab_results.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. exp_ab_results.json +160 -0
exp_ab_results.json ADDED
@@ -0,0 +1,160 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "experiment": "A_B_downstream_crosslingual",
3
+ "timestamp": "2026-04-13 09:59:30 UTC",
4
+ "base": {
5
+ "sentiment": {
6
+ "he": {
7
+ "logprob_acc": 0.53,
8
+ "logprob_n": 200,
9
+ "gen_acc": 0.0,
10
+ "gen_n": 50
11
+ },
12
+ "ar": {
13
+ "logprob_acc": 0.45,
14
+ "logprob_n": 200,
15
+ "gen_acc": 0.0,
16
+ "gen_n": 50
17
+ },
18
+ "fa": {
19
+ "logprob_acc": 0.605,
20
+ "logprob_n": 200,
21
+ "gen_acc": 0.0,
22
+ "gen_n": 50
23
+ },
24
+ "en": {
25
+ "logprob_acc": 0.515,
26
+ "logprob_n": 200,
27
+ "gen_acc": 0.0,
28
+ "gen_n": 50
29
+ }
30
+ },
31
+ "belebele": {
32
+ "he": {
33
+ "accuracy": 0.26,
34
+ "n": 100
35
+ },
36
+ "ar": {
37
+ "accuracy": 0.22,
38
+ "n": 100
39
+ },
40
+ "fa": {
41
+ "accuracy": 0.19,
42
+ "n": 100
43
+ },
44
+ "en": {
45
+ "accuracy": 0.19,
46
+ "n": 100
47
+ }
48
+ }
49
+ },
50
+ "en_sft": {
51
+ "sentiment": {
52
+ "he": {
53
+ "logprob_acc": 0.515,
54
+ "logprob_n": 200,
55
+ "gen_acc": 0.0,
56
+ "gen_n": 50
57
+ },
58
+ "ar": {
59
+ "logprob_acc": 0.465,
60
+ "logprob_n": 200,
61
+ "gen_acc": 0.0,
62
+ "gen_n": 50
63
+ },
64
+ "fa": {
65
+ "logprob_acc": 0.585,
66
+ "logprob_n": 200,
67
+ "gen_acc": 0.0,
68
+ "gen_n": 50
69
+ },
70
+ "en": {
71
+ "logprob_acc": 0.52,
72
+ "logprob_n": 200,
73
+ "gen_acc": 0.02,
74
+ "gen_n": 50
75
+ }
76
+ },
77
+ "belebele": {
78
+ "he": {
79
+ "accuracy": 0.26,
80
+ "n": 100
81
+ },
82
+ "ar": {
83
+ "accuracy": 0.23,
84
+ "n": 100
85
+ },
86
+ "fa": {
87
+ "accuracy": 0.2,
88
+ "n": 100
89
+ },
90
+ "en": {
91
+ "accuracy": 0.2,
92
+ "n": 100
93
+ }
94
+ }
95
+ },
96
+ "multilingual_sft": {
97
+ "sentiment": {
98
+ "he": {
99
+ "logprob_acc": 0.845,
100
+ "logprob_n": 200,
101
+ "gen_acc": 0.82,
102
+ "gen_n": 50
103
+ },
104
+ "ar": {
105
+ "logprob_acc": 0.605,
106
+ "logprob_n": 200,
107
+ "gen_acc": 0.64,
108
+ "gen_n": 50
109
+ },
110
+ "fa": {
111
+ "logprob_acc": 0.785,
112
+ "logprob_n": 200,
113
+ "gen_acc": 0.74,
114
+ "gen_n": 50
115
+ },
116
+ "en": {
117
+ "logprob_acc": 0.73,
118
+ "logprob_n": 200,
119
+ "gen_acc": 0.64,
120
+ "gen_n": 50
121
+ }
122
+ },
123
+ "belebele": {
124
+ "he": {
125
+ "accuracy": 0.28,
126
+ "n": 100
127
+ },
128
+ "ar": {
129
+ "accuracy": 0.24,
130
+ "n": 100
131
+ },
132
+ "fa": {
133
+ "accuracy": 0.25,
134
+ "n": 100
135
+ },
136
+ "en": {
137
+ "accuracy": 0.23,
138
+ "n": 100
139
+ }
140
+ }
141
+ },
142
+ "generation_samples": [
143
+ {
144
+ "prompt": "Write a short sentence in Hebrew about the weather.",
145
+ "response": "-הייתי עייף."
146
+ },
147
+ {
148
+ "prompt": "اكتب جملة قصيرة عن الطقس بالعربية.",
149
+ "response": "- I'm sorry."
150
+ },
151
+ {
152
+ "prompt": "یک جمله کوتاه درباره آب و هوا به فارسی بنویسید.",
153
+ "response": "- What is the weather like in the summer?"
154
+ },
155
+ {
156
+ "prompt": "What is the capital of Israel?",
157
+ "response": "מה זה?"
158
+ }
159
+ ]
160
+ }