Hugging Face's logo

wanlige
/
li-14b-v0.4

Text Generation
Transformers
Safetensors
qwen2
mergekit
Merge
conversational
Eval Results
text-generation-inference
Model card Files
xet
Community
3
wanlige commited on
Commit
976c15e
·
verified ·
1 Parent(s): 7fd12b8

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +21 -105
README.md CHANGED
@@ -12,195 +12,114 @@ base_model:
12
  - SicariusSicariiStuff/Impish_QWEN_14B-1M
13
  - Qwen/Qwen2.5-14B
14
  model-index:
15
-
16
  - name: li-14b-v0.4
17
-
18
  results:
19
-
20
  - task:
21
-
22
  type: text-generation
23
-
24
  name: Text Generation
25
-
26
  dataset:
27
-
28
  name: IFEval (0-Shot)
29
-
30
  type: HuggingFaceH4/ifeval
31
-
32
  args:
33
-
34
  num_few_shot: 0
35
-
36
  metrics:
37
-
38
  - type: inst_level_strict_acc and prompt_level_strict_acc
39
-
40
  value: 81.33
41
-
42
  name: strict accuracy
43
-
44
  source:
45
-
46
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
47
-
48
  name: Open LLM Leaderboard
49
-
50
  - task:
51
-
52
  type: text-generation
53
-
54
  name: Text Generation
55
-
56
  dataset:
57
-
58
  name: BBH (3-Shot)
59
-
60
  type: BBH
61
-
62
  args:
63
-
64
  num_few_shot: 3
65
-
66
  metrics:
67
-
68
  - type: acc_norm
69
-
70
  value: 50.38
71
-
72
  name: normalized accuracy
73
-
74
  source:
75
-
76
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
77
-
78
  name: Open LLM Leaderboard
79
-
80
  - task:
81
-
82
  type: text-generation
83
-
84
  name: Text Generation
85
-
86
  dataset:
87
-
88
  name: MATH Lvl 5 (4-Shot)
89
-
90
  type: hendrycks/competition_math
91
-
92
  args:
93
-
94
  num_few_shot: 4
95
-
96
  metrics:
97
-
98
  - type: exact_match
99
-
100
  value: 55.74
101
-
102
  name: exact match
103
-
104
  source:
105
-
106
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
107
-
108
  name: Open LLM Leaderboard
109
-
110
  - task:
111
-
112
  type: text-generation
113
-
114
  name: Text Generation
115
-
116
  dataset:
117
-
118
  name: GPQA (0-shot)
119
-
120
  type: Idavidrein/gpqa
121
-
122
  args:
123
-
124
  num_few_shot: 0
125
-
126
  metrics:
127
-
128
  - type: acc_norm
129
-
130
  value: 11.86
131
-
132
  name: acc_norm
133
-
134
  source:
135
-
136
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
137
-
138
  name: Open LLM Leaderboard
139
-
140
  - task:
141
-
142
  type: text-generation
143
-
144
  name: Text Generation
145
-
146
  dataset:
147
-
148
  name: MuSR (0-shot)
149
-
150
  type: TAUR-Lab/MuSR
151
-
152
  args:
153
-
154
  num_few_shot: 0
155
-
156
  metrics:
157
-
158
  - type: acc_norm
159
-
160
  value: 16.35
161
-
162
  name: acc_norm
163
-
164
  source:
165
-
166
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
167
-
168
  name: Open LLM Leaderboard
169
-
170
  - task:
171
-
172
  type: text-generation
173
-
174
  name: Text Generation
175
-
176
  dataset:
177
-
178
  name: MMLU-PRO (5-shot)
179
-
180
  type: TIGER-Lab/MMLU-Pro
181
-
182
  config: main
183
-
184
  split: test
185
-
186
  args:
187
-
188
  num_few_shot: 5
189
-
190
  metrics:
191
-
192
  - type: acc
193
-
194
  value: 46.3
195
-
196
  name: accuracy
197
-
198
  source:
199
-
200
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
201
-
202
  name: Open LLM Leaderboard
203
-
 
 
 
 
 
 
 
204
  ---
205
 
206
  > [!TIP] This model is currently ranked #1 among the models up to 15B parameters and #50 among all models on the [Open LLM Leaderboard](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard).
@@ -381,7 +300,4 @@ Detailed results can be found [here](https://huggingface.co/datasets/open-llm-le
381
 
382
  |MuSR (0-shot) |16.35|
383
 
384
- |MMLU-PRO (5-shot) |46.30|
385
-
386
-
387
-
 
12
  - SicariusSicariiStuff/Impish_QWEN_14B-1M
13
  - Qwen/Qwen2.5-14B
14
  model-index:
 
15
  - name: li-14b-v0.4
 
16
  results:
 
17
  - task:
 
18
  type: text-generation
 
19
  name: Text Generation
 
20
  dataset:
 
21
  name: IFEval (0-Shot)
 
22
  type: HuggingFaceH4/ifeval
 
23
  args:
 
24
  num_few_shot: 0
 
25
  metrics:
 
26
  - type: inst_level_strict_acc and prompt_level_strict_acc
 
27
  value: 81.33
 
28
  name: strict accuracy
 
29
  source:
30
+ url: >-
31
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
 
32
  name: Open LLM Leaderboard
 
33
  - task:
 
34
  type: text-generation
 
35
  name: Text Generation
 
36
  dataset:
 
37
  name: BBH (3-Shot)
 
38
  type: BBH
 
39
  args:
 
40
  num_few_shot: 3
 
41
  metrics:
 
42
  - type: acc_norm
 
43
  value: 50.38
 
44
  name: normalized accuracy
 
45
  source:
46
+ url: >-
47
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
 
48
  name: Open LLM Leaderboard
 
49
  - task:
 
50
  type: text-generation
 
51
  name: Text Generation
 
52
  dataset:
 
53
  name: MATH Lvl 5 (4-Shot)
 
54
  type: hendrycks/competition_math
 
55
  args:
 
56
  num_few_shot: 4
 
57
  metrics:
 
58
  - type: exact_match
 
59
  value: 55.74
 
60
  name: exact match
 
61
  source:
62
+ url: >-
63
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
 
64
  name: Open LLM Leaderboard
 
65
  - task:
 
66
  type: text-generation
 
67
  name: Text Generation
 
68
  dataset:
 
69
  name: GPQA (0-shot)
 
70
  type: Idavidrein/gpqa
 
71
  args:
 
72
  num_few_shot: 0
 
73
  metrics:
 
74
  - type: acc_norm
 
75
  value: 11.86
 
76
  name: acc_norm
 
77
  source:
78
+ url: >-
79
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
 
80
  name: Open LLM Leaderboard
 
81
  - task:
 
82
  type: text-generation
 
83
  name: Text Generation
 
84
  dataset:
 
85
  name: MuSR (0-shot)
 
86
  type: TAUR-Lab/MuSR
 
87
  args:
 
88
  num_few_shot: 0
 
89
  metrics:
 
90
  - type: acc_norm
 
91
  value: 16.35
 
92
  name: acc_norm
 
93
  source:
94
+ url: >-
95
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
 
96
  name: Open LLM Leaderboard
 
97
  - task:
 
98
  type: text-generation
 
99
  name: Text Generation
 
100
  dataset:
 
101
  name: MMLU-PRO (5-shot)
 
102
  type: TIGER-Lab/MMLU-Pro
 
103
  config: main
 
104
  split: test
 
105
  args:
 
106
  num_few_shot: 5
 
107
  metrics:
 
108
  - type: acc
 
109
  value: 46.3
 
110
  name: accuracy
 
111
  source:
112
+ url: >-
113
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=wanlige/li-14b-v0.4
 
114
  name: Open LLM Leaderboard
115
+ license: apache-2.0
116
+ language:
117
+ - zh
118
+ - en
119
+ - fr
120
+ - sa
121
+ - de
122
+ pipeline_tag: text-generation
123
  ---
124
 
125
  > [!TIP] This model is currently ranked #1 among the models up to 15B parameters and #50 among all models on the [Open LLM Leaderboard](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard).
 
300
 
301
  |MuSR (0-shot) |16.35|
302
 
303
+ |MMLU-PRO (5-shot) |46.30|