Add results of Llama 3 8b, Llama 3 70b, Qwen 2.5 72b (all Instruct fp16)

#3
by sm1rk - opened
data/BABILong NeurIPS24 Figs - leaderboard.csv CHANGED
@@ -1,171 +1,234 @@
1
- Model,task,0K,1K,2K,4K,8K,16K,32K,64K,128K,512K,1M,10M
2
- GPT-2 (137M),avg,27,15,,,,,,,,,,
3
- mamba-2.8b-hf,avg,70,52,35,9,0,,,,,,,
4
- rwkv-6-world-7b,avg,56,55,48,35,7,,,,,,,
5
- v5-Eagle-7B-HF,avg,62,54,48,41,2,,,,,,,
6
- Meta-Llama-3-8B-Instruct,avg,64,60,58,50,44,,,,,,,
7
- LLaMA-2-7B-32K,avg,41,53,45,40,39,32,3,,,,,
8
- longchat-7b-v1.5-32k,avg,46,42,40,41,42,39,5,,,,,
9
- LongAlpaca-13B,avg,48,47,46,43,40,36,4,,,,,
10
- Llama-2-7B-32K-Instruct,avg,49,52,49,43,40,35,5,,,,,
11
- 01-ai/Yi-34B,avg,72,52,43,37,38,31,4,,,,,
12
- Mistral-7b-Instruct-v0.2,avg,60,56,52,49,45,42,37,,,,,
13
- Mixtral-8x7B-Instruct-v0.1,avg,65,63,60,55,50,46,40,,,,,
14
- 01-ai/Yi-34B-200k,avg,65,59,56,54,52,50,48,48,,,,
15
- Mixtral-8x22B-Instruct-v0.1,avg,75,73,70,65,58,51,43,35,,,,
16
- activation-beacon-llama2-7b-chat,avg,55,52,47,43,36,23,16,8,6,,,
17
- Yarn-Mistral-7b-128k,avg,51,52,43,40,38,30,16,10,9,,,
18
- chatglm3-6b-128k,avg,56,55,51,48,46,41,36,21,13,,,
19
- activation-beacon-mistral-7b,avg,59,56,51,48,43,37,36,27,14,,,
20
- 01-ai/Yi-9B-200k,avg,52,55,48,46,45,36,37,29,24,,,
21
- Phi-3-mini-128k-instruct,avg,64,57,55,51,50,46,42,37,7,,,
22
- ai21labs/Jamba-v0.1,avg,65,53,50,48,46,45,41,40,34,,,
23
- c4ai-command-r-v01,avg,64,64,63,61,59,52,51,46,38,,,
24
- Meta-Llama-3.1-8B-Instruct,avg,67,68,66,66,62,60,56,49,39,,,
25
- Phi-3-medium-128k-instruct,avg,72,70,67,62,60,57,53,45,30,,,
26
- GPT-4,avg,87,81,77,74,71,64,53,43,36,,,
27
- Meta-Llama-3.1-70B-Instruct,avg,85,81,78,74,70,65,59,53,45,,,
28
- ~ Mamba (130M) fine-tune,avg,,,,"98,7","98,5","98,5","98,1",97,"92,5",,,
29
- Llama3-ChatQA-1.5-8B + RAG,avg,48,48,47,46,45,45,44,42,45,42,39,37
30
- ~ RMT (137M) fine-tune,avg,"99,36","97,4","94,66","92,32","89,9","85,62","77,88","69,86","58,52","46,36","42,84","33,78"
31
- ~ ARMT (137M) fine-tune,avg,"99,32",,,"98,1","98,2","98,1",98,"97,9","96,9","95,3","93,4","76,6"
32
- GPT-2 (137M),qa1,35,13,,,,,,,,,,
33
- mamba-2.8b-hf,qa1,65,56,40,7,1,,,,,,,
34
- rwkv-6-world-7b,qa1,65,62,53,39,5,,,,,,,
35
- v5-Eagle-7B-HF,qa1,68,58,52,36,3,,,,,,,
36
- Meta-Llama-3-8B-Instruct,qa1,98,93,90,79,62,,,,,,,
37
- LLaMA-2-7B-32K,qa1,54,57,33,26,34,32,3,,,,,
38
- longchat-7b-v1.5-32k,qa1,52,60,56,55,50,42,4,,,,,
39
- LongAlpaca-13B,qa1,58,55,58,50,37,23,2,,,,,
40
- Llama-2-7B-32K-Instruct,qa1,65,61,52,41,35,23,3,,,,,
41
- 01-ai/Yi-34B,qa1,99,59,51,34,46,31,4,,,,,
42
- Mistral-7b-Instruct-v0.2,qa1,92,86,75,64,63,57,45,,,,,
43
- Mixtral-8x7B-Instruct-v0.1,qa1,99,92,84,77,65,53,49,,,,,
44
- 01-ai/Yi-34B-200k,qa1,85,73,68,66,63,65,62,60,,,,
45
- Mixtral-8x22B-Instruct-v0.1,qa1,100,99,95,89,79,63,40,38,,,,
46
- activation-beacon-llama2-7b-chat,qa1,85,81,67,65,48,21,16,6,5,,,
47
- Yarn-Mistral-7b-128k,qa1,61,71,58,45,51,34,21,8,8,,,
48
- chatglm3-6b-128k,qa1,82,77,74,72,67,56,47,13,13,,,
49
- activation-beacon-mistral-7b,qa1,92,86,73,59,47,42,42,27,9,,,
50
- 01-ai/Yi-9B-200k,qa1,33,82,69,59,56,47,44,32,29,,,
51
- Phi-3-mini-128k-instruct,qa1,97,84,72,69,70,60,53,38,1,,,
52
- ai21labs/Jamba-v0.1,qa1,90,72,66,63,65,58,50,49,38,,,
53
- c4ai-command-r-v01,qa1,98,95,94,91,89,68,70,50,30,,,
54
- Phi-3-medium-128k-instruct,qa1,100,93,87,80,81,72,69,58,21,,,
55
- GPT-4,qa1,99,100,100,95,93,84,62,58,39,,,
56
- ~ Mamba (130M) fine-tune,qa1,100,100,100,100,100,100,100,100,100,92,,
57
- Llama3-ChatQA-1.5-8B + RAG,qa1,60,62,60,58,58,60,60,56,64,54,55,50
58
- ~ RMT (137M) fine-tune,qa1,100,100,100,100,100,100,99,96,94,87,84,66
59
- ~ ARMT (137M) fine-tune,qa1,100,,,100,100,100,100,100,100,99,99,89
60
- GPT-2 (137M),qa2,21,17,,,,,,,,,,
61
- mamba-2.8b-hf,qa2,68,47,24,8,0,,,,,,,
62
- rwkv-6-world-7b,qa2,42,26,20,16,3,,,,,,,
63
- v5-Eagle-7B-HF,qa2,43,27,25,19,2,,,,,,,
64
- Meta-Llama-3-8B-Instruct,qa2,47,46,49,39,20,,,,,,,
65
- LLaMA-2-7B-32K,qa2,37,40,37,26,27,21,2,,,,,
66
- longchat-7b-v1.5-32k,qa2,31,19,14,13,18,14,2,,,,,
67
- LongAlpaca-13B,qa2,30,30,28,29,26,18,2,,,,,
68
- Llama-2-7B-32K-Instruct,qa2,41,40,36,27,20,16,3,,,,,
69
- 01-ai/Yi-34B,qa2,67,43,32,30,23,15,4,,,,,
70
- Mistral-7b-Instruct-v0.2,qa2,46,34,32,22,11,9,7,,,,,
71
- Mixtral-8x7B-Instruct-v0.1,qa2,58,51,46,35,27,18,14,,,,,
72
- 01-ai/Yi-34B-200k,qa2,62,51,46,42,35,32,27,30,,,,
73
- Mixtral-8x22B-Instruct-v0.1,qa2,77,65,61,56,48,41,33,11,,,,
74
- activation-beacon-llama2-7b-chat,qa2,52,37,28,25,23,11,9,4,2,,,
75
- Yarn-Mistral-7b-128k,qa2,47,48,37,30,34,21,12,5,3,,,
76
- chatglm3-6b-128k,qa2,51,46,45,39,38,31,24,6,5,,,
77
- activation-beacon-mistral-7b,qa2,45,35,32,28,22,14,12,10,2,,,
78
- 01-ai/Yi-9B-200k,qa2,67,52,43,39,31,25,22,12,8,,,
79
- Phi-3-mini-128k-instruct,qa2,57,38,38,36,34,23,22,15,2,,,
80
- ai21labs/Jamba-v0.1,qa2,57,43,42,39,37,29,26,20,16,,,
81
- c4ai-command-r-v01,qa2,64,58,56,54,50,39,37,32,16,,,
82
- Phi-3-medium-128k-instruct,qa2,76,62,58,51,44,41,27,14,11,,,
83
- GPT-4,qa2,88,79,72,68,65,59,42,25,25,,,
84
- ~ Mamba (130M) fine-tune,qa2,98,98,98,98,98,98,98,95,87,,,
85
- Llama3-ChatQA-1.5-8B + RAG,qa2,28,25,22,19,14,13,9,7,6,6,2,2
86
- ~ RMT (137M) fine-tune,qa2,100,100,99,98,97,94,82,59,39,25,22,19
87
- ~ ARMT (137M) fine-tune,qa2,100,,,100,100,100,100,100,100,99,99,84
88
- GPT-2 (137M),qa3,6,8,,,,,,,,,,
89
- mamba-2.8b-hf,qa3,48,39,21,8,0,,,,,,,
90
- rwkv-6-world-7b,qa3,40,45,28,24,4,,,,,,,
91
- v5-Eagle-7B-HF,qa3,43,34,30,40,1,,,,,,,
92
- Meta-Llama-3-8B-Instruct,qa3,33,28,30,26,11,,,,,,,
93
- LLaMA-2-7B-32K,qa3,32,38,34,28,27,21,1,,,,,
94
- longchat-7b-v1.5-32k,qa3,22,16,15,17,21,22,4,,,,,
95
- LongAlpaca-13B,qa3,25,26,26,25,24,27,2,,,,,
96
- Llama-2-7B-32K-Instruct,qa3,35,36,34,26,23,20,2,,,,,
97
- 01-ai/Yi-34B,qa3,45,34,24,20,17,12,4,,,,,
98
- Mistral-7b-Instruct-v0.2,qa3,36,34,31,30,24,18,12,,,,,
99
- Mixtral-8x7B-Instruct-v0.1,qa3,34,32,31,30,27,29,24,,,,,
100
- 01-ai/Yi-34B-200k,qa3,35,30,27,24,24,22,22,26,,,,
101
- Mixtral-8x22B-Instruct-v0.1,qa3,53,56,49,39,31,27,26,26,,,,
102
- activation-beacon-llama2-7b-chat,qa3,33,25,25,21,20,17,13,5,5,,,
103
- Yarn-Mistral-7b-128k,qa3,31,36,33,32,27,25,9,13,7,,,
104
- chatglm3-6b-128k,qa3,33,37,31,31,27,25,23,17,9,,,
105
- activation-beacon-mistral-7b,qa3,36,33,25,21,18,15,15,15,16,,,
106
- 01-ai/Yi-9B-200k,qa3,34,33,29,24,25,21,20,20,8,,,
107
- Phi-3-mini-128k-instruct,qa3,32,41,31,27,26,24,21,22,4,,,
108
- ai21labs/Jamba-v0.1,qa3,32,31,29,26,24,22,22,21,26,,,
109
- c4ai-command-r-v01,qa3,25,28,26,28,26,30,28,33,24,,,
110
- Phi-3-medium-128k-instruct,qa3,53,51,45,35,30,30,27,25,17,,,
111
- GPT-4,qa3,56,63,57,56,53,45,31,31,32,,,
112
- ~ Mamba (130M) fine-tune,qa3,97,97,97,97,97,96,95,92,81,,,
113
- Llama3-ChatQA-1.5-8B + RAG,qa3,17,18,17,17,16,17,15,13,19,17,10,11
114
- ~ RMT (137M) fine-tune,qa3,97,94,88,81,73,66,55,55,36,25,22,21
115
- ~ ARMT (137M) fine-tune,qa3,97,,,92,92,92,91,90,86,80,72,37
116
- GPT-2 (137M),qa4,29,18,,,,,,,,,,
117
- mamba-2.8b-hf,qa4,96,59,47,12,0,,,,,,,
118
- rwkv-6-world-7b,qa4,54,65,57,35,7,,,,,,,
119
- v5-Eagle-7B-HF,qa4,79,74,63,55,3,,,,,,,
120
- Meta-Llama-3-8B-Instruct,qa4,58,55,50,43,52,,,,,,,
121
- LLaMA-2-7B-32K,qa4,26,54,51,51,46,36,3,,,,,
122
- longchat-7b-v1.5-32k,qa4,60,55,52,57,57,49,4,,,,,
123
- LongAlpaca-13B,qa4,65,61,58,52,50,44,4,,,,,
124
- Llama-2-7B-32K-Instruct,qa4,39,52,54,56,55,52,6,,,,,
125
- 01-ai/Yi-34B,qa4,59,56,51,55,52,43,4,,,,,
126
- Mistral-7b-Instruct-v0.2,qa4,54,58,58,60,60,58,54,,,,,
127
- Mixtral-8x7B-Instruct-v0.1,qa4,55,60,59,61,63,61,58,,,,,
128
- 01-ai/Yi-34B-200k,qa4,64,65,64,63,61,56,54,44,,,,
129
- Mixtral-8x22B-Instruct-v0.1,qa4,56,62,59,62,62,60,54,39,,,,
130
- activation-beacon-llama2-7b-chat,qa4,40,50,52,43,34,22,14,9,10,,,
131
- Yarn-Mistral-7b-128k,qa4,60,56,43,45,32,31,16,7,8,,,
132
- chatglm3-6b-128k,qa4,45,48,42,38,35,32,27,13,11,,,
133
- activation-beacon-mistral-7b,qa4,53,58,60,57,53,50,45,29,15,,,
134
- 01-ai/Yi-9B-200k,qa4,49,47,50,50,54,43,45,36,33,,,
135
- Phi-3-mini-128k-instruct,qa4,54,56,56,50,49,50,45,47,5,,,
136
- ai21labs/Jamba-v0.1,qa4,64,50,49,49,48,52,46,49,38,,,
137
- c4ai-command-r-v01,qa4,46,58,59,54,56,46,46,47,52,,,
138
- Phi-3-medium-128k-instruct,qa4,54,61,63,64,64,61,59,52,33,,,
139
- GPT-4,qa4,98,70,63,60,52,47,46,40,32,,,
140
- ~ Mamba (130M) fine-tune,qa4,100,100,100,100,100,100,99,100,98,,,
141
- Llama3-ChatQA-1.5-8B + RAG,qa4,53,58,56,59,57,60,60,59,60,59,54,56
142
- ~ RMT (137M) fine-tune,qa4,100,94,87,83,80,75,64,51,38,26,24,20
143
- ~ ARMT (137M) fine-tune,qa4,100,,,100,100,100,100,100,100,100,100,92
144
- GPT-2 (137M),qa5,45,19,,,,,,,,,,
145
- mamba-2.8b-hf,qa5,75,58,43,9,0,,,,,,,
146
- rwkv-6-world-7b,qa5,79,77,80,61,14,,,,,,,
147
- v5-Eagle-7B-HF,qa5,75,76,71,57,3,,,,,,,
148
- Meta-Llama-3-8B-Instruct,qa5,85,78,73,65,73,,,,,,,
149
- LLaMA-2-7B-32K,qa5,55,74,70,67,59,51,7,,,,,
150
- longchat-7b-v1.5-32k,qa5,63,62,62,65,66,67,9,,,,,
151
- LongAlpaca-13B,qa5,63,61,61,61,62,66,12,,,,,
152
- Llama-2-7B-32K-Instruct,qa5,63,69,69,67,66,63,9,,,,,
153
- 01-ai/Yi-34B,qa5,88,70,59,48,53,55,4,,,,,
154
- Mistral-7b-Instruct-v0.2,qa5,70,66,66,67,69,67,67,,,,,
155
- Mixtral-8x7B-Instruct-v0.1,qa5,80,79,80,73,66,67,56,,,,,
156
- 01-ai/Yi-34B-200k,qa5,78,77,77,76,76,75,76,80,,,,
157
- Mixtral-8x22B-Instruct-v0.1,qa5,87,84,84,79,69,64,63,63,,,,
158
- activation-beacon-llama2-7b-chat,qa5,65,67,64,63,57,45,29,17,7,,,
159
- Yarn-Mistral-7b-128k,qa5,58,47,45,47,47,38,23,17,19,,,
160
- chatglm3-6b-128k,qa5,70,69,64,60,61,61,58,55,26,,,
161
- activation-beacon-mistral-7b,qa5,68,66,66,74,74,66,67,55,28,,,
162
- 01-ai/Yi-9B-200k,qa5,76,59,50,57,57,45,52,47,40,,,
163
- Phi-3-mini-128k-instruct,qa5,79,66,76,72,72,73,71,64,23,,,
164
- ai21labs/Jamba-v0.1,qa5,83,70,64,62,58,64,63,60,50,,,
165
- c4ai-command-r-v01,qa5,86,82,81,78,75,79,72,70,66,,,
166
- Phi-3-medium-128k-instruct,qa5,77,85,84,81,82,82,81,78,69,,,
167
- GPT-4,qa5,96,95,92,90,93,85,82,60,51,,,
168
- ~ Mamba (130M) fine-tune,qa5,98,99,98,99,99,99,98,99,98,,,
169
- Llama3-ChatQA-1.5-8B + RAG,qa5,80,77,78,77,78,77,78,76,75,75,76,67
170
- ~ RMT (137M) fine-tune,qa5,100,100,99,99,99,94,90,89,86,69,63,44
171
- ~ ARMT (137M) fine-tune,qa5,"99,6",,,"98,1","98,2","98,1",98,"97,9","96,9","95,3","93,4","76,6"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Model,task,0K,1K,2K,4K,8K,16K,32K,64K,128K,512K,1M,10M
2
+ GPT-2 (137M),avg,27,15,,,,,,,,,,
3
+ mamba-2.8b-hf,avg,70,52,35,9,0,,,,,,,
4
+ rwkv-6-world-7b,avg,56,55,48,35,7,,,,,,,
5
+ v5-Eagle-7B-HF,avg,62,54,48,41,2,,,,,,,
6
+ Meta-Llama-3-8B-Instruct,avg,64,60,58,50,44,,,,,,,
7
+ LLaMA-2-7B-32K,avg,41,53,45,40,39,32,3,,,,,
8
+ longchat-7b-v1.5-32k,avg,46,42,40,41,42,39,5,,,,,
9
+ LongAlpaca-13B,avg,48,47,46,43,40,36,4,,,,,
10
+ Llama-2-7B-32K-Instruct,avg,49,52,49,43,40,35,5,,,,,
11
+ 01-ai/Yi-34B,avg,72,52,43,37,38,31,4,,,,,
12
+ Mistral-7b-Instruct-v0.2,avg,60,56,52,49,45,42,37,,,,,
13
+ Mixtral-8x7B-Instruct-v0.1,avg,65,63,60,55,50,46,40,,,,,
14
+ 01-ai/Yi-34B-200k,avg,65,59,56,54,52,50,48,48,,,,
15
+ Mixtral-8x22B-Instruct-v0.1,avg,75,73,70,65,58,51,43,35,,,,
16
+ activation-beacon-llama2-7b-chat,avg,55,52,47,43,36,23,16,8,6,,,
17
+ Yarn-Mistral-7b-128k,avg,51,52,43,40,38,30,16,10,9,,,
18
+ chatglm3-6b-128k,avg,56,55,51,48,46,41,36,21,13,,,
19
+ activation-beacon-mistral-7b,avg,59,56,51,48,43,37,36,27,14,,,
20
+ 01-ai/Yi-9B-200k,avg,52,55,48,46,45,36,37,29,24,,,
21
+ Phi-3-mini-128k-instruct,avg,64,57,55,51,50,46,42,37,7,,,
22
+ ai21labs/Jamba-v0.1,avg,65,53,50,48,46,45,41,40,34,,,
23
+ c4ai-command-r-v01,avg,64,64,63,61,59,52,51,46,38,,,
24
+ Meta-Llama-3.1-8B-Instruct,avg,67,68,66,66,62,60,56,49,39,,,
25
+ Phi-3-medium-128k-instruct,avg,72,70,67,62,60,57,53,45,30,,,
26
+ GPT-4,avg,87,81,77,74,71,64,53,43,36,,,
27
+ Meta-Llama-3.1-70B-Instruct,avg,85,81,78,74,70,65,59,53,45,,,
28
+ ~ Mamba (130M) fine-tune,avg,,,,"98,7","98,5","98,5","98,1",97,"92,5",,,
29
+ Llama3-ChatQA-1.5-8B + RAG,avg,48,48,47,46,45,45,44,42,45,42,39,37
30
+ ~ RMT (137M) fine-tune,avg,"99,36","97,4","94,66","92,32","89,9","85,62","77,88","69,86","58,52","46,36","42,84","33,78"
31
+ ~ ARMT (137M) fine-tune,avg,"99,32",,,"98,1","98,2","98,1",98,"97,9","96,9","95,3","93,4","76,6"
32
+ GPT-2 (137M),qa1,35,13,,,,,,,,,,
33
+ mamba-2.8b-hf,qa1,65,56,40,7,1,,,,,,,
34
+ rwkv-6-world-7b,qa1,65,62,53,39,5,,,,,,,
35
+ v5-Eagle-7B-HF,qa1,68,58,52,36,3,,,,,,,
36
+ Meta-Llama-3-8B-Instruct,qa1,98,93,90,79,62,,,,,,,
37
+ LLaMA-2-7B-32K,qa1,54,57,33,26,34,32,3,,,,,
38
+ longchat-7b-v1.5-32k,qa1,52,60,56,55,50,42,4,,,,,
39
+ LongAlpaca-13B,qa1,58,55,58,50,37,23,2,,,,,
40
+ Llama-2-7B-32K-Instruct,qa1,65,61,52,41,35,23,3,,,,,
41
+ 01-ai/Yi-34B,qa1,99,59,51,34,46,31,4,,,,,
42
+ Mistral-7b-Instruct-v0.2,qa1,92,86,75,64,63,57,45,,,,,
43
+ Mixtral-8x7B-Instruct-v0.1,qa1,99,92,84,77,65,53,49,,,,,
44
+ 01-ai/Yi-34B-200k,qa1,85,73,68,66,63,65,62,60,,,,
45
+ Mixtral-8x22B-Instruct-v0.1,qa1,100,99,95,89,79,63,40,38,,,,
46
+ activation-beacon-llama2-7b-chat,qa1,85,81,67,65,48,21,16,6,5,,,
47
+ Yarn-Mistral-7b-128k,qa1,61,71,58,45,51,34,21,8,8,,,
48
+ chatglm3-6b-128k,qa1,82,77,74,72,67,56,47,13,13,,,
49
+ activation-beacon-mistral-7b,qa1,92,86,73,59,47,42,42,27,9,,,
50
+ 01-ai/Yi-9B-200k,qa1,33,82,69,59,56,47,44,32,29,,,
51
+ Phi-3-mini-128k-instruct,qa1,97,84,72,69,70,60,53,38,1,,,
52
+ ai21labs/Jamba-v0.1,qa1,90,72,66,63,65,58,50,49,38,,,
53
+ c4ai-command-r-v01,qa1,98,95,94,91,89,68,70,50,30,,,
54
+ Phi-3-medium-128k-instruct,qa1,100,93,87,80,81,72,69,58,21,,,
55
+ GPT-4,qa1,99,100,100,95,93,84,62,58,39,,,
56
+ ~ Mamba (130M) fine-tune,qa1,100,100,100,100,100,100,100,100,100,92,,
57
+ Llama3-ChatQA-1.5-8B + RAG,qa1,60,62,60,58,58,60,60,56,64,54,55,50
58
+ ~ RMT (137M) fine-tune,qa1,100,100,100,100,100,100,99,96,94,87,84,66
59
+ ~ ARMT (137M) fine-tune,qa1,100,,,100,100,100,100,100,100,99,99,89
60
+ GPT-2 (137M),qa2,21,17,,,,,,,,,,
61
+ mamba-2.8b-hf,qa2,68,47,24,8,0,,,,,,,
62
+ rwkv-6-world-7b,qa2,42,26,20,16,3,,,,,,,
63
+ v5-Eagle-7B-HF,qa2,43,27,25,19,2,,,,,,,
64
+ Meta-Llama-3-8B-Instruct,qa2,47,46,49,39,20,,,,,,,
65
+ LLaMA-2-7B-32K,qa2,37,40,37,26,27,21,2,,,,,
66
+ longchat-7b-v1.5-32k,qa2,31,19,14,13,18,14,2,,,,,
67
+ LongAlpaca-13B,qa2,30,30,28,29,26,18,2,,,,,
68
+ Llama-2-7B-32K-Instruct,qa2,41,40,36,27,20,16,3,,,,,
69
+ 01-ai/Yi-34B,qa2,67,43,32,30,23,15,4,,,,,
70
+ Mistral-7b-Instruct-v0.2,qa2,46,34,32,22,11,9,7,,,,,
71
+ Mixtral-8x7B-Instruct-v0.1,qa2,58,51,46,35,27,18,14,,,,,
72
+ 01-ai/Yi-34B-200k,qa2,62,51,46,42,35,32,27,30,,,,
73
+ Mixtral-8x22B-Instruct-v0.1,qa2,77,65,61,56,48,41,33,11,,,,
74
+ activation-beacon-llama2-7b-chat,qa2,52,37,28,25,23,11,9,4,2,,,
75
+ Yarn-Mistral-7b-128k,qa2,47,48,37,30,34,21,12,5,3,,,
76
+ chatglm3-6b-128k,qa2,51,46,45,39,38,31,24,6,5,,,
77
+ activation-beacon-mistral-7b,qa2,45,35,32,28,22,14,12,10,2,,,
78
+ 01-ai/Yi-9B-200k,qa2,67,52,43,39,31,25,22,12,8,,,
79
+ Phi-3-mini-128k-instruct,qa2,57,38,38,36,34,23,22,15,2,,,
80
+ ai21labs/Jamba-v0.1,qa2,57,43,42,39,37,29,26,20,16,,,
81
+ c4ai-command-r-v01,qa2,64,58,56,54,50,39,37,32,16,,,
82
+ Phi-3-medium-128k-instruct,qa2,76,62,58,51,44,41,27,14,11,,,
83
+ GPT-4,qa2,88,79,72,68,65,59,42,25,25,,,
84
+ ~ Mamba (130M) fine-tune,qa2,98,98,98,98,98,98,98,95,87,,,
85
+ Llama3-ChatQA-1.5-8B + RAG,qa2,28,25,22,19,14,13,9,7,6,6,2,2
86
+ ~ RMT (137M) fine-tune,qa2,100,100,99,98,97,94,82,59,39,25,22,19
87
+ ~ ARMT (137M) fine-tune,qa2,100,,,100,100,100,100,100,100,99,99,84
88
+ GPT-2 (137M),qa3,6,8,,,,,,,,,,
89
+ mamba-2.8b-hf,qa3,48,39,21,8,0,,,,,,,
90
+ rwkv-6-world-7b,qa3,40,45,28,24,4,,,,,,,
91
+ v5-Eagle-7B-HF,qa3,43,34,30,40,1,,,,,,,
92
+ Meta-Llama-3-8B-Instruct,qa3,33,28,30,26,11,,,,,,,
93
+ LLaMA-2-7B-32K,qa3,32,38,34,28,27,21,1,,,,,
94
+ longchat-7b-v1.5-32k,qa3,22,16,15,17,21,22,4,,,,,
95
+ LongAlpaca-13B,qa3,25,26,26,25,24,27,2,,,,,
96
+ Llama-2-7B-32K-Instruct,qa3,35,36,34,26,23,20,2,,,,,
97
+ 01-ai/Yi-34B,qa3,45,34,24,20,17,12,4,,,,,
98
+ Mistral-7b-Instruct-v0.2,qa3,36,34,31,30,24,18,12,,,,,
99
+ Mixtral-8x7B-Instruct-v0.1,qa3,34,32,31,30,27,29,24,,,,,
100
+ 01-ai/Yi-34B-200k,qa3,35,30,27,24,24,22,22,26,,,,
101
+ Mixtral-8x22B-Instruct-v0.1,qa3,53,56,49,39,31,27,26,26,,,,
102
+ activation-beacon-llama2-7b-chat,qa3,33,25,25,21,20,17,13,5,5,,,
103
+ Yarn-Mistral-7b-128k,qa3,31,36,33,32,27,25,9,13,7,,,
104
+ chatglm3-6b-128k,qa3,33,37,31,31,27,25,23,17,9,,,
105
+ activation-beacon-mistral-7b,qa3,36,33,25,21,18,15,15,15,16,,,
106
+ 01-ai/Yi-9B-200k,qa3,34,33,29,24,25,21,20,20,8,,,
107
+ Phi-3-mini-128k-instruct,qa3,32,41,31,27,26,24,21,22,4,,,
108
+ ai21labs/Jamba-v0.1,qa3,32,31,29,26,24,22,22,21,26,,,
109
+ c4ai-command-r-v01,qa3,25,28,26,28,26,30,28,33,24,,,
110
+ Phi-3-medium-128k-instruct,qa3,53,51,45,35,30,30,27,25,17,,,
111
+ GPT-4,qa3,56,63,57,56,53,45,31,31,32,,,
112
+ ~ Mamba (130M) fine-tune,qa3,97,97,97,97,97,96,95,92,81,,,
113
+ Llama3-ChatQA-1.5-8B + RAG,qa3,17,18,17,17,16,17,15,13,19,17,10,11
114
+ ~ RMT (137M) fine-tune,qa3,97,94,88,81,73,66,55,55,36,25,22,21
115
+ ~ ARMT (137M) fine-tune,qa3,97,,,92,92,92,91,90,86,80,72,37
116
+ GPT-2 (137M),qa4,29,18,,,,,,,,,,
117
+ mamba-2.8b-hf,qa4,96,59,47,12,0,,,,,,,
118
+ rwkv-6-world-7b,qa4,54,65,57,35,7,,,,,,,
119
+ v5-Eagle-7B-HF,qa4,79,74,63,55,3,,,,,,,
120
+ Meta-Llama-3-8B-Instruct,qa4,58,55,50,43,52,,,,,,,
121
+ LLaMA-2-7B-32K,qa4,26,54,51,51,46,36,3,,,,,
122
+ longchat-7b-v1.5-32k,qa4,60,55,52,57,57,49,4,,,,,
123
+ LongAlpaca-13B,qa4,65,61,58,52,50,44,4,,,,,
124
+ Llama-2-7B-32K-Instruct,qa4,39,52,54,56,55,52,6,,,,,
125
+ 01-ai/Yi-34B,qa4,59,56,51,55,52,43,4,,,,,
126
+ Mistral-7b-Instruct-v0.2,qa4,54,58,58,60,60,58,54,,,,,
127
+ Mixtral-8x7B-Instruct-v0.1,qa4,55,60,59,61,63,61,58,,,,,
128
+ 01-ai/Yi-34B-200k,qa4,64,65,64,63,61,56,54,44,,,,
129
+ Mixtral-8x22B-Instruct-v0.1,qa4,56,62,59,62,62,60,54,39,,,,
130
+ activation-beacon-llama2-7b-chat,qa4,40,50,52,43,34,22,14,9,10,,,
131
+ Yarn-Mistral-7b-128k,qa4,60,56,43,45,32,31,16,7,8,,,
132
+ chatglm3-6b-128k,qa4,45,48,42,38,35,32,27,13,11,,,
133
+ activation-beacon-mistral-7b,qa4,53,58,60,57,53,50,45,29,15,,,
134
+ 01-ai/Yi-9B-200k,qa4,49,47,50,50,54,43,45,36,33,,,
135
+ Phi-3-mini-128k-instruct,qa4,54,56,56,50,49,50,45,47,5,,,
136
+ ai21labs/Jamba-v0.1,qa4,64,50,49,49,48,52,46,49,38,,,
137
+ c4ai-command-r-v01,qa4,46,58,59,54,56,46,46,47,52,,,
138
+ Phi-3-medium-128k-instruct,qa4,54,61,63,64,64,61,59,52,33,,,
139
+ GPT-4,qa4,98,70,63,60,52,47,46,40,32,,,
140
+ ~ Mamba (130M) fine-tune,qa4,100,100,100,100,100,100,99,100,98,,,
141
+ Llama3-ChatQA-1.5-8B + RAG,qa4,53,58,56,59,57,60,60,59,60,59,54,56
142
+ ~ RMT (137M) fine-tune,qa4,100,94,87,83,80,75,64,51,38,26,24,20
143
+ ~ ARMT (137M) fine-tune,qa4,100,,,100,100,100,100,100,100,100,100,92
144
+ GPT-2 (137M),qa5,45,19,,,,,,,,,,
145
+ mamba-2.8b-hf,qa5,75,58,43,9,0,,,,,,,
146
+ rwkv-6-world-7b,qa5,79,77,80,61,14,,,,,,,
147
+ v5-Eagle-7B-HF,qa5,75,76,71,57,3,,,,,,,
148
+ Meta-Llama-3-8B-Instruct,qa5,85,78,73,65,73,,,,,,,
149
+ LLaMA-2-7B-32K,qa5,55,74,70,67,59,51,7,,,,,
150
+ longchat-7b-v1.5-32k,qa5,63,62,62,65,66,67,9,,,,,
151
+ LongAlpaca-13B,qa5,63,61,61,61,62,66,12,,,,,
152
+ Llama-2-7B-32K-Instruct,qa5,63,69,69,67,66,63,9,,,,,
153
+ 01-ai/Yi-34B,qa5,88,70,59,48,53,55,4,,,,,
154
+ Mistral-7b-Instruct-v0.2,qa5,70,66,66,67,69,67,67,,,,,
155
+ Mixtral-8x7B-Instruct-v0.1,qa5,80,79,80,73,66,67,56,,,,,
156
+ 01-ai/Yi-34B-200k,qa5,78,77,77,76,76,75,76,80,,,,
157
+ Mixtral-8x22B-Instruct-v0.1,qa5,87,84,84,79,69,64,63,63,,,,
158
+ activation-beacon-llama2-7b-chat,qa5,65,67,64,63,57,45,29,17,7,,,
159
+ Yarn-Mistral-7b-128k,qa5,58,47,45,47,47,38,23,17,19,,,
160
+ chatglm3-6b-128k,qa5,70,69,64,60,61,61,58,55,26,,,
161
+ activation-beacon-mistral-7b,qa5,68,66,66,74,74,66,67,55,28,,,
162
+ 01-ai/Yi-9B-200k,qa5,76,59,50,57,57,45,52,47,40,,,
163
+ Phi-3-mini-128k-instruct,qa5,79,66,76,72,72,73,71,64,23,,,
164
+ ai21labs/Jamba-v0.1,qa5,83,70,64,62,58,64,63,60,50,,,
165
+ c4ai-command-r-v01,qa5,86,82,81,78,75,79,72,70,66,,,
166
+ Phi-3-medium-128k-instruct,qa5,77,85,84,81,82,82,81,78,69,,,
167
+ GPT-4,qa5,96,95,92,90,93,85,82,60,51,,,
168
+ ~ Mamba (130M) fine-tune,qa5,98,99,98,99,99,99,98,99,98,,,
169
+ Llama3-ChatQA-1.5-8B + RAG,qa5,80,77,78,77,78,77,78,76,75,75,76,67
170
+ ~ RMT (137M) fine-tune,qa5,100,100,99,99,99,94,90,89,86,69,63,44
171
+ ~ ARMT (137M) fine-tune,qa5,"99,6",,,"98,1","98,2","98,1",98,"97,9","96,9","95,3","93,4","76,6"
172
+ llama3.1:8b-instruct-fp16,qa1,86,,,70,67,,,,,,,,
173
+ llama3.1:70b-instruct-fp16,qa1,99,,,79,69,,,,,,,,
174
+ qwen2.5:72b-instruct-fp16,qa1,100,,,90,89,,,,,,,,
175
+ llama3.1:8b-instruct-fp16,qa2,47,,,35,37,,,,,,,,
176
+ llama3.1:70b-instruct-fp16,qa2,71,,,43,39,,,,,,,,
177
+ qwen2.5:72b-instruct-fp16,qa2,77,,,53,53,,,,,,,,
178
+ llama3.1:8b-instruct-fp16,qa3,27,,,28,28,,,,,,,,
179
+ llama3.1:70b-instruct-fp16,qa3,39,,,24,24,,,,,,,,
180
+ qwen2.5:72b-instruct-fp16,qa3,34,,,40,30,,,,,,,,
181
+ llama3.1:8b-instruct-fp16,qa4,69,,,56,54,,,,,,,,
182
+ llama3.1:70b-instruct-fp16,qa4,91,,,64,66,,,,,,,,
183
+ qwen2.5:72b-instruct-fp16,qa4,95,,,64,57,,,,,,,,
184
+ llama3.1:8b-instruct-fp16,qa5,81,,,73,71,,,,,,,,
185
+ llama3.1:70b-instruct-fp16,qa5,86,,,81,78,,,,,,,,
186
+ qwen2.5:72b-instruct-fp16,qa5,88,,,86,89,,,,,,,,
187
+ llama3.1:8b-instruct-fp16,qa6,95,,,78,70,,,,,,,,
188
+ llama3.1:70b-instruct-fp16,qa6,97,,,90,92,,,,,,,,
189
+ qwen2.5:72b-instruct-fp16,qa6,97,,,93,91,,,,,,,,
190
+ llama3.1:8b-instruct-fp16,qa7,46,,,12,12,,,,,,,,
191
+ llama3.1:70b-instruct-fp16,qa7,92,,,47,38,,,,,,,,
192
+ qwen2.5:72b-instruct-fp16,qa7,95,,,59,56,,,,,,,,
193
+ llama3.1:8b-instruct-fp16,qa8,39,,,25,15,,,,,,,,
194
+ llama3.1:70b-instruct-fp16,qa8,91,,,47,36,,,,,,,,
195
+ qwen2.5:72b-instruct-fp16,qa8,94,,,66,64,,,,,,,,
196
+ llama3.1:8b-instruct-fp16,qa9,95,,,86,79,,,,,,,,
197
+ llama3.1:70b-instruct-fp16,qa9,97,,,96,93,,,,,,,,
198
+ qwen2.5:72b-instruct-fp16,qa9,99,,,96,94,,,,,,,,
199
+ llama3.1:8b-instruct-fp16,qa10,77,,,65,63,,,,,,,,
200
+ llama3.1:70b-instruct-fp16,qa10,86,,,84,76,,,,,,,,
201
+ qwen2.5:72b-instruct-fp16,qa10,89,,,89,91,,,,,,,,
202
+ llama3.1:8b-instruct-fp16,qa11,88,,,,,,,,,,
203
+ llama3.1:70b-instruct-fp16,qa11,100,,,,,,,,,,
204
+ qwen2.5:72b-instruct-fp16,qa11,100,,,,,,,,,,
205
+ llama3.1:8b-instruct-fp16,qa12,76,,,,,,,,,,
206
+ llama3.1:70b-instruct-fp16,qa12,100,,,,,,,,,,
207
+ qwen2.5:72b-instruct-fp16,qa12,95,,,,,,,,,,
208
+ llama3.1:8b-instruct-fp16,qa13,67,,,,,,,,,,
209
+ llama3.1:70b-instruct-fp16,qa13,99,,,,,,,,,,
210
+ qwen2.5:72b-instruct-fp16,qa13,100,,,,,,,,,,
211
+ llama3.1:8b-instruct-fp16,qa14,86,,,,,,,,,,
212
+ llama3.1:70b-instruct-fp16,qa14,86,,,,,,,,,,
213
+ qwen2.5:72b-instruct-fp16,qa14,90,,,,,,,,,,
214
+ llama3.1:8b-instruct-fp16,qa15,78,,,,,,,,,,
215
+ llama3.1:70b-instruct-fp16,qa15,95,,,,,,,,,,
216
+ qwen2.5:72b-instruct-fp16,qa15,83,,,,,,,,,,
217
+ llama3.1:8b-instruct-fp16,qa16,63,,,,,,,,,,
218
+ llama3.1:70b-instruct-fp16,qa16,78,,,,,,,,,,
219
+ qwen2.5:72b-instruct-fp16,qa16,92,,,,,,,,,,
220
+ llama3.1:8b-instruct-fp16,qa17,51,,,,,,,,,,
221
+ llama3.1:70b-instruct-fp16,qa17,79,,,,,,,,,,
222
+ qwen2.5:72b-instruct-fp16,qa17,77,,,,,,,,,,
223
+ llama3.1:8b-instruct-fp16,qa18,76,,,,,,,,,,
224
+ llama3.1:70b-instruct-fp16,qa18,97,,,,,,,,,,
225
+ qwen2.5:72b-instruct-fp16,qa18,98,,,,,,,,,,
226
+ llama3.1:8b-instruct-fp16,qa19,11,,,,,,,,,,
227
+ llama3.1:70b-instruct-fp16,qa19,51,,,,,,,,,,
228
+ qwen2.5:72b-instruct-fp16,qa19,34,,,,,,,,,,
229
+ llama3.1:8b-instruct-fp16,qa20,45,,,,,,,,,,
230
+ llama3.1:70b-instruct-fp16,qa20,73,,,,,,,,,,
231
+ qwen2.5:72b-instruct-fp16,qa20,56,,,,,,,,,,
232
+ llama3.1:8b-instruct-fp16,avg,65,,,59,55,,,,,,,,
233
+ llama3.1:70b-instruct-fp16,avg,85,,,74,68,,,,,,,,
234
+ qwen2.5:72b-instruct-fp16,avg,84,,,82,79,,,,,,,,