SaylorTwift HF staff commited on
Commit
878567c
1 Parent(s): bc77b55

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +68 -542
README.md CHANGED
@@ -3,8 +3,8 @@ pretty_name: Evaluation run of lmsys/vicuna-7b-v1.5-16k
3
  dataset_summary: "Dataset automatically created during the evaluation run of model\
4
  \ [lmsys/vicuna-7b-v1.5-16k](https://huggingface.co/lmsys/vicuna-7b-v1.5-16k) on\
5
  \ the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard).\n\
6
- \nThe dataset is composed of 61 configuration, each one coresponding to one of the\
7
- \ evaluated task.\n\nThe dataset has been created from 1 run(s). Each run can be\
8
  \ found as a specific split in each configuration, the split being named using the\
9
  \ timestamp of the run.The \"train\" split is always pointing to the latest results.\n\
10
  \nAn additional configuration \"results\" store all the aggregated results of the\
@@ -12,173 +12,20 @@ dataset_summary: "Dataset automatically created during the evaluation run of mod
12
  \ Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)).\n\
13
  \nTo load the details from a run, you can for instance do the following:\n```python\n\
14
  from datasets import load_dataset\ndata = load_dataset(\"open-llm-leaderboard/details_lmsys__vicuna-7b-v1.5-16k\"\
15
- ,\n\t\"harness_truthfulqa_mc_0\",\n\tsplit=\"train\")\n```\n\n## Latest results\n\
16
- \nThese are the [latest results from run 2023-08-18T07:58:23.659880](https://huggingface.co/datasets/open-llm-leaderboard/details_lmsys__vicuna-7b-v1.5-16k/blob/main/results_2023-08-18T07%3A58%3A23.659880.json)\
17
- \ (note that their might be results for other tasks in the repos if successive evals\
18
- \ didn't cover the same tasks. You find each in the results and the \"latest\" split\
19
- \ for each eval):\n\n```python\n{\n \"all\": {\n \"acc\": 0.4968600651239149,\n\
20
- \ \"acc_stderr\": 0.035128371186062546,\n \"acc_norm\": 0.500705169805845,\n\
21
- \ \"acc_norm_stderr\": 0.035114818289212105,\n \"mc1\": 0.33659730722154224,\n\
22
- \ \"mc1_stderr\": 0.016542412809494887,\n \"mc2\": 0.5041278188644791,\n\
23
- \ \"mc2_stderr\": 0.01576728095387478\n },\n \"harness|arc:challenge|25\"\
24
- : {\n \"acc\": 0.507679180887372,\n \"acc_stderr\": 0.01460966744089257,\n\
25
- \ \"acc_norm\": 0.5469283276450512,\n \"acc_norm_stderr\": 0.014546892052005628\n\
26
- \ },\n \"harness|hellaswag|10\": {\n \"acc\": 0.5855407289384584,\n\
27
- \ \"acc_stderr\": 0.004916216503770338,\n \"acc_norm\": 0.7731527584146585,\n\
28
- \ \"acc_norm_stderr\": 0.0041793709784810045\n },\n \"harness|hendrycksTest-abstract_algebra|5\"\
29
- : {\n \"acc\": 0.28,\n \"acc_stderr\": 0.04512608598542128,\n \
30
- \ \"acc_norm\": 0.28,\n \"acc_norm_stderr\": 0.04512608598542128\n \
31
- \ },\n \"harness|hendrycksTest-anatomy|5\": {\n \"acc\": 0.4222222222222222,\n\
32
- \ \"acc_stderr\": 0.04266763404099582,\n \"acc_norm\": 0.4222222222222222,\n\
33
- \ \"acc_norm_stderr\": 0.04266763404099582\n },\n \"harness|hendrycksTest-astronomy|5\"\
34
- : {\n \"acc\": 0.4934210526315789,\n \"acc_stderr\": 0.040685900502249704,\n\
35
- \ \"acc_norm\": 0.4934210526315789,\n \"acc_norm_stderr\": 0.040685900502249704\n\
36
- \ },\n \"harness|hendrycksTest-business_ethics|5\": {\n \"acc\": 0.54,\n\
37
- \ \"acc_stderr\": 0.05009082659620333,\n \"acc_norm\": 0.54,\n \
38
- \ \"acc_norm_stderr\": 0.05009082659620333\n },\n \"harness|hendrycksTest-clinical_knowledge|5\"\
39
- : {\n \"acc\": 0.5132075471698113,\n \"acc_stderr\": 0.030762134874500476,\n\
40
- \ \"acc_norm\": 0.5132075471698113,\n \"acc_norm_stderr\": 0.030762134874500476\n\
41
- \ },\n \"harness|hendrycksTest-college_biology|5\": {\n \"acc\": 0.5,\n\
42
- \ \"acc_stderr\": 0.04181210050035455,\n \"acc_norm\": 0.5,\n \
43
- \ \"acc_norm_stderr\": 0.04181210050035455\n },\n \"harness|hendrycksTest-college_chemistry|5\"\
44
- : {\n \"acc\": 0.28,\n \"acc_stderr\": 0.04512608598542127,\n \
45
- \ \"acc_norm\": 0.28,\n \"acc_norm_stderr\": 0.04512608598542127\n \
46
- \ },\n \"harness|hendrycksTest-college_computer_science|5\": {\n \"acc\"\
47
- : 0.42,\n \"acc_stderr\": 0.049604496374885836,\n \"acc_norm\": 0.42,\n\
48
- \ \"acc_norm_stderr\": 0.049604496374885836\n },\n \"harness|hendrycksTest-college_mathematics|5\"\
49
- : {\n \"acc\": 0.27,\n \"acc_stderr\": 0.044619604333847394,\n \
50
- \ \"acc_norm\": 0.27,\n \"acc_norm_stderr\": 0.044619604333847394\n \
51
- \ },\n \"harness|hendrycksTest-college_medicine|5\": {\n \"acc\": 0.4277456647398844,\n\
52
- \ \"acc_stderr\": 0.03772446857518027,\n \"acc_norm\": 0.4277456647398844,\n\
53
- \ \"acc_norm_stderr\": 0.03772446857518027\n },\n \"harness|hendrycksTest-college_physics|5\"\
54
- : {\n \"acc\": 0.2549019607843137,\n \"acc_stderr\": 0.043364327079931785,\n\
55
- \ \"acc_norm\": 0.2549019607843137,\n \"acc_norm_stderr\": 0.043364327079931785\n\
56
- \ },\n \"harness|hendrycksTest-computer_security|5\": {\n \"acc\":\
57
- \ 0.59,\n \"acc_stderr\": 0.049431107042371025,\n \"acc_norm\": 0.59,\n\
58
- \ \"acc_norm_stderr\": 0.049431107042371025\n },\n \"harness|hendrycksTest-conceptual_physics|5\"\
59
- : {\n \"acc\": 0.425531914893617,\n \"acc_stderr\": 0.03232146916224468,\n\
60
- \ \"acc_norm\": 0.425531914893617,\n \"acc_norm_stderr\": 0.03232146916224468\n\
61
- \ },\n \"harness|hendrycksTest-econometrics|5\": {\n \"acc\": 0.40350877192982454,\n\
62
- \ \"acc_stderr\": 0.046151869625837026,\n \"acc_norm\": 0.40350877192982454,\n\
63
- \ \"acc_norm_stderr\": 0.046151869625837026\n },\n \"harness|hendrycksTest-electrical_engineering|5\"\
64
- : {\n \"acc\": 0.4827586206896552,\n \"acc_stderr\": 0.04164188720169377,\n\
65
- \ \"acc_norm\": 0.4827586206896552,\n \"acc_norm_stderr\": 0.04164188720169377\n\
66
- \ },\n \"harness|hendrycksTest-elementary_mathematics|5\": {\n \"acc\"\
67
- : 0.31746031746031744,\n \"acc_stderr\": 0.02397386199899206,\n \"\
68
- acc_norm\": 0.31746031746031744,\n \"acc_norm_stderr\": 0.02397386199899206\n\
69
- \ },\n \"harness|hendrycksTest-formal_logic|5\": {\n \"acc\": 0.30952380952380953,\n\
70
- \ \"acc_stderr\": 0.04134913018303316,\n \"acc_norm\": 0.30952380952380953,\n\
71
- \ \"acc_norm_stderr\": 0.04134913018303316\n },\n \"harness|hendrycksTest-global_facts|5\"\
72
- : {\n \"acc\": 0.33,\n \"acc_stderr\": 0.047258156262526045,\n \
73
- \ \"acc_norm\": 0.33,\n \"acc_norm_stderr\": 0.047258156262526045\n \
74
- \ },\n \"harness|hendrycksTest-high_school_biology|5\": {\n \"acc\"\
75
- : 0.5483870967741935,\n \"acc_stderr\": 0.028310500348568392,\n \"\
76
- acc_norm\": 0.5483870967741935,\n \"acc_norm_stderr\": 0.028310500348568392\n\
77
- \ },\n \"harness|hendrycksTest-high_school_chemistry|5\": {\n \"acc\"\
78
- : 0.3399014778325123,\n \"acc_stderr\": 0.033327690684107895,\n \"\
79
- acc_norm\": 0.3399014778325123,\n \"acc_norm_stderr\": 0.033327690684107895\n\
80
- \ },\n \"harness|hendrycksTest-high_school_computer_science|5\": {\n \
81
- \ \"acc\": 0.48,\n \"acc_stderr\": 0.050211673156867795,\n \"acc_norm\"\
82
- : 0.48,\n \"acc_norm_stderr\": 0.050211673156867795\n },\n \"harness|hendrycksTest-high_school_european_history|5\"\
83
- : {\n \"acc\": 0.6484848484848484,\n \"acc_stderr\": 0.0372820699868265,\n\
84
- \ \"acc_norm\": 0.6484848484848484,\n \"acc_norm_stderr\": 0.0372820699868265\n\
85
- \ },\n \"harness|hendrycksTest-high_school_geography|5\": {\n \"acc\"\
86
- : 0.6111111111111112,\n \"acc_stderr\": 0.0347327959083696,\n \"acc_norm\"\
87
- : 0.6111111111111112,\n \"acc_norm_stderr\": 0.0347327959083696\n },\n\
88
- \ \"harness|hendrycksTest-high_school_government_and_politics|5\": {\n \
89
- \ \"acc\": 0.694300518134715,\n \"acc_stderr\": 0.033248379397581594,\n\
90
- \ \"acc_norm\": 0.694300518134715,\n \"acc_norm_stderr\": 0.033248379397581594\n\
91
- \ },\n \"harness|hendrycksTest-high_school_macroeconomics|5\": {\n \
92
- \ \"acc\": 0.45384615384615384,\n \"acc_stderr\": 0.02524277098712618,\n\
93
- \ \"acc_norm\": 0.45384615384615384,\n \"acc_norm_stderr\": 0.02524277098712618\n\
94
- \ },\n \"harness|hendrycksTest-high_school_mathematics|5\": {\n \"\
95
- acc\": 0.2851851851851852,\n \"acc_stderr\": 0.027528599210340496,\n \
96
- \ \"acc_norm\": 0.2851851851851852,\n \"acc_norm_stderr\": 0.027528599210340496\n\
97
- \ },\n \"harness|hendrycksTest-high_school_microeconomics|5\": {\n \
98
- \ \"acc\": 0.4495798319327731,\n \"acc_stderr\": 0.03231293497137707,\n \
99
- \ \"acc_norm\": 0.4495798319327731,\n \"acc_norm_stderr\": 0.03231293497137707\n\
100
- \ },\n \"harness|hendrycksTest-high_school_physics|5\": {\n \"acc\"\
101
- : 0.2781456953642384,\n \"acc_stderr\": 0.03658603262763743,\n \"\
102
- acc_norm\": 0.2781456953642384,\n \"acc_norm_stderr\": 0.03658603262763743\n\
103
- \ },\n \"harness|hendrycksTest-high_school_psychology|5\": {\n \"acc\"\
104
- : 0.6770642201834862,\n \"acc_stderr\": 0.020048115923415318,\n \"\
105
- acc_norm\": 0.6770642201834862,\n \"acc_norm_stderr\": 0.020048115923415318\n\
106
- \ },\n \"harness|hendrycksTest-high_school_statistics|5\": {\n \"acc\"\
107
- : 0.3611111111111111,\n \"acc_stderr\": 0.03275773486100999,\n \"\
108
- acc_norm\": 0.3611111111111111,\n \"acc_norm_stderr\": 0.03275773486100999\n\
109
- \ },\n \"harness|hendrycksTest-high_school_us_history|5\": {\n \"acc\"\
110
- : 0.6519607843137255,\n \"acc_stderr\": 0.03343311240488419,\n \"\
111
- acc_norm\": 0.6519607843137255,\n \"acc_norm_stderr\": 0.03343311240488419\n\
112
- \ },\n \"harness|hendrycksTest-high_school_world_history|5\": {\n \"\
113
- acc\": 0.7130801687763713,\n \"acc_stderr\": 0.029443773022594693,\n \
114
- \ \"acc_norm\": 0.7130801687763713,\n \"acc_norm_stderr\": 0.029443773022594693\n\
115
- \ },\n \"harness|hendrycksTest-human_aging|5\": {\n \"acc\": 0.6188340807174888,\n\
116
- \ \"acc_stderr\": 0.03259625118416828,\n \"acc_norm\": 0.6188340807174888,\n\
117
- \ \"acc_norm_stderr\": 0.03259625118416828\n },\n \"harness|hendrycksTest-human_sexuality|5\"\
118
- : {\n \"acc\": 0.5725190839694656,\n \"acc_stderr\": 0.04338920305792401,\n\
119
- \ \"acc_norm\": 0.5725190839694656,\n \"acc_norm_stderr\": 0.04338920305792401\n\
120
- \ },\n \"harness|hendrycksTest-international_law|5\": {\n \"acc\":\
121
- \ 0.6033057851239669,\n \"acc_stderr\": 0.044658697805310094,\n \"\
122
- acc_norm\": 0.6033057851239669,\n \"acc_norm_stderr\": 0.044658697805310094\n\
123
- \ },\n \"harness|hendrycksTest-jurisprudence|5\": {\n \"acc\": 0.5555555555555556,\n\
124
- \ \"acc_stderr\": 0.04803752235190192,\n \"acc_norm\": 0.5555555555555556,\n\
125
- \ \"acc_norm_stderr\": 0.04803752235190192\n },\n \"harness|hendrycksTest-logical_fallacies|5\"\
126
- : {\n \"acc\": 0.5460122699386503,\n \"acc_stderr\": 0.0391170190467718,\n\
127
- \ \"acc_norm\": 0.5460122699386503,\n \"acc_norm_stderr\": 0.0391170190467718\n\
128
- \ },\n \"harness|hendrycksTest-machine_learning|5\": {\n \"acc\": 0.38392857142857145,\n\
129
- \ \"acc_stderr\": 0.04616143075028547,\n \"acc_norm\": 0.38392857142857145,\n\
130
- \ \"acc_norm_stderr\": 0.04616143075028547\n },\n \"harness|hendrycksTest-management|5\"\
131
- : {\n \"acc\": 0.6504854368932039,\n \"acc_stderr\": 0.047211885060971716,\n\
132
- \ \"acc_norm\": 0.6504854368932039,\n \"acc_norm_stderr\": 0.047211885060971716\n\
133
- \ },\n \"harness|hendrycksTest-marketing|5\": {\n \"acc\": 0.7435897435897436,\n\
134
- \ \"acc_stderr\": 0.028605953702004243,\n \"acc_norm\": 0.7435897435897436,\n\
135
- \ \"acc_norm_stderr\": 0.028605953702004243\n },\n \"harness|hendrycksTest-medical_genetics|5\"\
136
- : {\n \"acc\": 0.57,\n \"acc_stderr\": 0.049756985195624284,\n \
137
- \ \"acc_norm\": 0.57,\n \"acc_norm_stderr\": 0.049756985195624284\n \
138
- \ },\n \"harness|hendrycksTest-miscellaneous|5\": {\n \"acc\": 0.6551724137931034,\n\
139
- \ \"acc_stderr\": 0.016997123346113443,\n \"acc_norm\": 0.6551724137931034,\n\
140
- \ \"acc_norm_stderr\": 0.016997123346113443\n },\n \"harness|hendrycksTest-moral_disputes|5\"\
141
- : {\n \"acc\": 0.5346820809248555,\n \"acc_stderr\": 0.02685425792825887,\n\
142
- \ \"acc_norm\": 0.5346820809248555,\n \"acc_norm_stderr\": 0.02685425792825887\n\
143
- \ },\n \"harness|hendrycksTest-moral_scenarios|5\": {\n \"acc\": 0.24804469273743016,\n\
144
- \ \"acc_stderr\": 0.01444415780826144,\n \"acc_norm\": 0.24804469273743016,\n\
145
- \ \"acc_norm_stderr\": 0.01444415780826144\n },\n \"harness|hendrycksTest-nutrition|5\"\
146
- : {\n \"acc\": 0.5784313725490197,\n \"acc_stderr\": 0.028275490156791462,\n\
147
- \ \"acc_norm\": 0.5784313725490197,\n \"acc_norm_stderr\": 0.028275490156791462\n\
148
- \ },\n \"harness|hendrycksTest-philosophy|5\": {\n \"acc\": 0.5401929260450161,\n\
149
- \ \"acc_stderr\": 0.028306190403305696,\n \"acc_norm\": 0.5401929260450161,\n\
150
- \ \"acc_norm_stderr\": 0.028306190403305696\n },\n \"harness|hendrycksTest-prehistory|5\"\
151
- : {\n \"acc\": 0.5462962962962963,\n \"acc_stderr\": 0.0277012284685426,\n\
152
- \ \"acc_norm\": 0.5462962962962963,\n \"acc_norm_stderr\": 0.0277012284685426\n\
153
- \ },\n \"harness|hendrycksTest-professional_accounting|5\": {\n \"\
154
- acc\": 0.3971631205673759,\n \"acc_stderr\": 0.029189805673587105,\n \
155
- \ \"acc_norm\": 0.3971631205673759,\n \"acc_norm_stderr\": 0.029189805673587105\n\
156
- \ },\n \"harness|hendrycksTest-professional_law|5\": {\n \"acc\": 0.35919165580182527,\n\
157
- \ \"acc_stderr\": 0.012253386187584253,\n \"acc_norm\": 0.35919165580182527,\n\
158
- \ \"acc_norm_stderr\": 0.012253386187584253\n },\n \"harness|hendrycksTest-professional_medicine|5\"\
159
- : {\n \"acc\": 0.4852941176470588,\n \"acc_stderr\": 0.03035969707904611,\n\
160
- \ \"acc_norm\": 0.4852941176470588,\n \"acc_norm_stderr\": 0.03035969707904611\n\
161
- \ },\n \"harness|hendrycksTest-professional_psychology|5\": {\n \"\
162
- acc\": 0.48366013071895425,\n \"acc_stderr\": 0.020217030653186457,\n \
163
- \ \"acc_norm\": 0.48366013071895425,\n \"acc_norm_stderr\": 0.020217030653186457\n\
164
- \ },\n \"harness|hendrycksTest-public_relations|5\": {\n \"acc\": 0.6,\n\
165
- \ \"acc_stderr\": 0.0469237132203465,\n \"acc_norm\": 0.6,\n \
166
- \ \"acc_norm_stderr\": 0.0469237132203465\n },\n \"harness|hendrycksTest-security_studies|5\"\
167
- : {\n \"acc\": 0.5918367346938775,\n \"acc_stderr\": 0.03146465712827424,\n\
168
- \ \"acc_norm\": 0.5918367346938775,\n \"acc_norm_stderr\": 0.03146465712827424\n\
169
- \ },\n \"harness|hendrycksTest-sociology|5\": {\n \"acc\": 0.6467661691542289,\n\
170
- \ \"acc_stderr\": 0.03379790611796777,\n \"acc_norm\": 0.6467661691542289,\n\
171
- \ \"acc_norm_stderr\": 0.03379790611796777\n },\n \"harness|hendrycksTest-us_foreign_policy|5\"\
172
- : {\n \"acc\": 0.74,\n \"acc_stderr\": 0.04408440022768079,\n \
173
- \ \"acc_norm\": 0.74,\n \"acc_norm_stderr\": 0.04408440022768079\n \
174
- \ },\n \"harness|hendrycksTest-virology|5\": {\n \"acc\": 0.4397590361445783,\n\
175
- \ \"acc_stderr\": 0.03864139923699121,\n \"acc_norm\": 0.4397590361445783,\n\
176
- \ \"acc_norm_stderr\": 0.03864139923699121\n },\n \"harness|hendrycksTest-world_religions|5\"\
177
- : {\n \"acc\": 0.6783625730994152,\n \"acc_stderr\": 0.03582529442573122,\n\
178
- \ \"acc_norm\": 0.6783625730994152,\n \"acc_norm_stderr\": 0.03582529442573122\n\
179
- \ },\n \"harness|truthfulqa:mc|0\": {\n \"mc1\": 0.33659730722154224,\n\
180
- \ \"mc1_stderr\": 0.016542412809494887,\n \"mc2\": 0.5041278188644791,\n\
181
- \ \"mc2_stderr\": 0.01576728095387478\n }\n}\n```"
182
  repo_url: https://huggingface.co/lmsys/vicuna-7b-v1.5-16k
183
  leaderboard_url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard
184
  point_of_contact: clementine@hf.co
@@ -191,6 +38,22 @@ configs:
191
  - split: latest
192
  path:
193
  - '**/details_harness|arc:challenge|25_2023-08-18T07:58:23.659880.parquet'
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
194
  - config_name: harness_hellaswag_10
195
  data_files:
196
  - split: 2023_08_18T07_58_23.659880
@@ -783,14 +646,25 @@ configs:
783
  - split: latest
784
  path:
785
  - '**/details_harness|truthfulqa:mc|0_2023-08-18T07:58:23.659880.parquet'
 
 
 
 
 
 
 
 
786
  - config_name: results
787
  data_files:
788
  - split: 2023_08_18T07_58_23.659880
789
  path:
790
  - results_2023-08-18T07:58:23.659880.parquet
 
 
 
791
  - split: latest
792
  path:
793
- - results_2023-08-18T07:58:23.659880.parquet
794
  ---
795
 
796
  # Dataset Card for Evaluation run of lmsys/vicuna-7b-v1.5-16k
@@ -807,9 +681,9 @@ configs:
807
 
808
  Dataset automatically created during the evaluation run of model [lmsys/vicuna-7b-v1.5-16k](https://huggingface.co/lmsys/vicuna-7b-v1.5-16k) on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard).
809
 
810
- The dataset is composed of 61 configuration, each one coresponding to one of the evaluated task.
811
 
812
- The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
813
 
814
  An additional configuration "results" store all the aggregated results of the run (and is used to compute and display the agregated metrics on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)).
815
 
@@ -817,385 +691,37 @@ To load the details from a run, you can for instance do the following:
817
  ```python
818
  from datasets import load_dataset
819
  data = load_dataset("open-llm-leaderboard/details_lmsys__vicuna-7b-v1.5-16k",
820
- "harness_truthfulqa_mc_0",
821
  split="train")
822
  ```
823
 
824
  ## Latest results
825
 
826
- These are the [latest results from run 2023-08-18T07:58:23.659880](https://huggingface.co/datasets/open-llm-leaderboard/details_lmsys__vicuna-7b-v1.5-16k/blob/main/results_2023-08-18T07%3A58%3A23.659880.json) (note that their might be results for other tasks in the repos if successive evals didn't cover the same tasks. You find each in the results and the "latest" split for each eval):
827
 
828
  ```python
829
  {
830
  "all": {
831
- "acc": 0.4968600651239149,
832
- "acc_stderr": 0.035128371186062546,
833
- "acc_norm": 0.500705169805845,
834
- "acc_norm_stderr": 0.035114818289212105,
835
- "mc1": 0.33659730722154224,
836
- "mc1_stderr": 0.016542412809494887,
837
- "mc2": 0.5041278188644791,
838
- "mc2_stderr": 0.01576728095387478
839
- },
840
- "harness|arc:challenge|25": {
841
- "acc": 0.507679180887372,
842
- "acc_stderr": 0.01460966744089257,
843
- "acc_norm": 0.5469283276450512,
844
- "acc_norm_stderr": 0.014546892052005628
845
- },
846
- "harness|hellaswag|10": {
847
- "acc": 0.5855407289384584,
848
- "acc_stderr": 0.004916216503770338,
849
- "acc_norm": 0.7731527584146585,
850
- "acc_norm_stderr": 0.0041793709784810045
851
- },
852
- "harness|hendrycksTest-abstract_algebra|5": {
853
- "acc": 0.28,
854
- "acc_stderr": 0.04512608598542128,
855
- "acc_norm": 0.28,
856
- "acc_norm_stderr": 0.04512608598542128
857
- },
858
- "harness|hendrycksTest-anatomy|5": {
859
- "acc": 0.4222222222222222,
860
- "acc_stderr": 0.04266763404099582,
861
- "acc_norm": 0.4222222222222222,
862
- "acc_norm_stderr": 0.04266763404099582
863
- },
864
- "harness|hendrycksTest-astronomy|5": {
865
- "acc": 0.4934210526315789,
866
- "acc_stderr": 0.040685900502249704,
867
- "acc_norm": 0.4934210526315789,
868
- "acc_norm_stderr": 0.040685900502249704
869
- },
870
- "harness|hendrycksTest-business_ethics|5": {
871
- "acc": 0.54,
872
- "acc_stderr": 0.05009082659620333,
873
- "acc_norm": 0.54,
874
- "acc_norm_stderr": 0.05009082659620333
875
- },
876
- "harness|hendrycksTest-clinical_knowledge|5": {
877
- "acc": 0.5132075471698113,
878
- "acc_stderr": 0.030762134874500476,
879
- "acc_norm": 0.5132075471698113,
880
- "acc_norm_stderr": 0.030762134874500476
881
- },
882
- "harness|hendrycksTest-college_biology|5": {
883
- "acc": 0.5,
884
- "acc_stderr": 0.04181210050035455,
885
- "acc_norm": 0.5,
886
- "acc_norm_stderr": 0.04181210050035455
887
- },
888
- "harness|hendrycksTest-college_chemistry|5": {
889
- "acc": 0.28,
890
- "acc_stderr": 0.04512608598542127,
891
- "acc_norm": 0.28,
892
- "acc_norm_stderr": 0.04512608598542127
893
- },
894
- "harness|hendrycksTest-college_computer_science|5": {
895
- "acc": 0.42,
896
- "acc_stderr": 0.049604496374885836,
897
- "acc_norm": 0.42,
898
- "acc_norm_stderr": 0.049604496374885836
899
- },
900
- "harness|hendrycksTest-college_mathematics|5": {
901
- "acc": 0.27,
902
- "acc_stderr": 0.044619604333847394,
903
- "acc_norm": 0.27,
904
- "acc_norm_stderr": 0.044619604333847394
905
- },
906
- "harness|hendrycksTest-college_medicine|5": {
907
- "acc": 0.4277456647398844,
908
- "acc_stderr": 0.03772446857518027,
909
- "acc_norm": 0.4277456647398844,
910
- "acc_norm_stderr": 0.03772446857518027
911
- },
912
- "harness|hendrycksTest-college_physics|5": {
913
- "acc": 0.2549019607843137,
914
- "acc_stderr": 0.043364327079931785,
915
- "acc_norm": 0.2549019607843137,
916
- "acc_norm_stderr": 0.043364327079931785
917
- },
918
- "harness|hendrycksTest-computer_security|5": {
919
- "acc": 0.59,
920
- "acc_stderr": 0.049431107042371025,
921
- "acc_norm": 0.59,
922
- "acc_norm_stderr": 0.049431107042371025
923
- },
924
- "harness|hendrycksTest-conceptual_physics|5": {
925
- "acc": 0.425531914893617,
926
- "acc_stderr": 0.03232146916224468,
927
- "acc_norm": 0.425531914893617,
928
- "acc_norm_stderr": 0.03232146916224468
929
- },
930
- "harness|hendrycksTest-econometrics|5": {
931
- "acc": 0.40350877192982454,
932
- "acc_stderr": 0.046151869625837026,
933
- "acc_norm": 0.40350877192982454,
934
- "acc_norm_stderr": 0.046151869625837026
935
- },
936
- "harness|hendrycksTest-electrical_engineering|5": {
937
- "acc": 0.4827586206896552,
938
- "acc_stderr": 0.04164188720169377,
939
- "acc_norm": 0.4827586206896552,
940
- "acc_norm_stderr": 0.04164188720169377
941
- },
942
- "harness|hendrycksTest-elementary_mathematics|5": {
943
- "acc": 0.31746031746031744,
944
- "acc_stderr": 0.02397386199899206,
945
- "acc_norm": 0.31746031746031744,
946
- "acc_norm_stderr": 0.02397386199899206
947
- },
948
- "harness|hendrycksTest-formal_logic|5": {
949
- "acc": 0.30952380952380953,
950
- "acc_stderr": 0.04134913018303316,
951
- "acc_norm": 0.30952380952380953,
952
- "acc_norm_stderr": 0.04134913018303316
953
- },
954
- "harness|hendrycksTest-global_facts|5": {
955
- "acc": 0.33,
956
- "acc_stderr": 0.047258156262526045,
957
- "acc_norm": 0.33,
958
- "acc_norm_stderr": 0.047258156262526045
959
- },
960
- "harness|hendrycksTest-high_school_biology|5": {
961
- "acc": 0.5483870967741935,
962
- "acc_stderr": 0.028310500348568392,
963
- "acc_norm": 0.5483870967741935,
964
- "acc_norm_stderr": 0.028310500348568392
965
- },
966
- "harness|hendrycksTest-high_school_chemistry|5": {
967
- "acc": 0.3399014778325123,
968
- "acc_stderr": 0.033327690684107895,
969
- "acc_norm": 0.3399014778325123,
970
- "acc_norm_stderr": 0.033327690684107895
971
- },
972
- "harness|hendrycksTest-high_school_computer_science|5": {
973
- "acc": 0.48,
974
- "acc_stderr": 0.050211673156867795,
975
- "acc_norm": 0.48,
976
- "acc_norm_stderr": 0.050211673156867795
977
- },
978
- "harness|hendrycksTest-high_school_european_history|5": {
979
- "acc": 0.6484848484848484,
980
- "acc_stderr": 0.0372820699868265,
981
- "acc_norm": 0.6484848484848484,
982
- "acc_norm_stderr": 0.0372820699868265
983
- },
984
- "harness|hendrycksTest-high_school_geography|5": {
985
- "acc": 0.6111111111111112,
986
- "acc_stderr": 0.0347327959083696,
987
- "acc_norm": 0.6111111111111112,
988
- "acc_norm_stderr": 0.0347327959083696
989
- },
990
- "harness|hendrycksTest-high_school_government_and_politics|5": {
991
- "acc": 0.694300518134715,
992
- "acc_stderr": 0.033248379397581594,
993
- "acc_norm": 0.694300518134715,
994
- "acc_norm_stderr": 0.033248379397581594
995
- },
996
- "harness|hendrycksTest-high_school_macroeconomics|5": {
997
- "acc": 0.45384615384615384,
998
- "acc_stderr": 0.02524277098712618,
999
- "acc_norm": 0.45384615384615384,
1000
- "acc_norm_stderr": 0.02524277098712618
1001
- },
1002
- "harness|hendrycksTest-high_school_mathematics|5": {
1003
- "acc": 0.2851851851851852,
1004
- "acc_stderr": 0.027528599210340496,
1005
- "acc_norm": 0.2851851851851852,
1006
- "acc_norm_stderr": 0.027528599210340496
1007
- },
1008
- "harness|hendrycksTest-high_school_microeconomics|5": {
1009
- "acc": 0.4495798319327731,
1010
- "acc_stderr": 0.03231293497137707,
1011
- "acc_norm": 0.4495798319327731,
1012
- "acc_norm_stderr": 0.03231293497137707
1013
- },
1014
- "harness|hendrycksTest-high_school_physics|5": {
1015
- "acc": 0.2781456953642384,
1016
- "acc_stderr": 0.03658603262763743,
1017
- "acc_norm": 0.2781456953642384,
1018
- "acc_norm_stderr": 0.03658603262763743
1019
- },
1020
- "harness|hendrycksTest-high_school_psychology|5": {
1021
- "acc": 0.6770642201834862,
1022
- "acc_stderr": 0.020048115923415318,
1023
- "acc_norm": 0.6770642201834862,
1024
- "acc_norm_stderr": 0.020048115923415318
1025
- },
1026
- "harness|hendrycksTest-high_school_statistics|5": {
1027
- "acc": 0.3611111111111111,
1028
- "acc_stderr": 0.03275773486100999,
1029
- "acc_norm": 0.3611111111111111,
1030
- "acc_norm_stderr": 0.03275773486100999
1031
- },
1032
- "harness|hendrycksTest-high_school_us_history|5": {
1033
- "acc": 0.6519607843137255,
1034
- "acc_stderr": 0.03343311240488419,
1035
- "acc_norm": 0.6519607843137255,
1036
- "acc_norm_stderr": 0.03343311240488419
1037
- },
1038
- "harness|hendrycksTest-high_school_world_history|5": {
1039
- "acc": 0.7130801687763713,
1040
- "acc_stderr": 0.029443773022594693,
1041
- "acc_norm": 0.7130801687763713,
1042
- "acc_norm_stderr": 0.029443773022594693
1043
- },
1044
- "harness|hendrycksTest-human_aging|5": {
1045
- "acc": 0.6188340807174888,
1046
- "acc_stderr": 0.03259625118416828,
1047
- "acc_norm": 0.6188340807174888,
1048
- "acc_norm_stderr": 0.03259625118416828
1049
- },
1050
- "harness|hendrycksTest-human_sexuality|5": {
1051
- "acc": 0.5725190839694656,
1052
- "acc_stderr": 0.04338920305792401,
1053
- "acc_norm": 0.5725190839694656,
1054
- "acc_norm_stderr": 0.04338920305792401
1055
- },
1056
- "harness|hendrycksTest-international_law|5": {
1057
- "acc": 0.6033057851239669,
1058
- "acc_stderr": 0.044658697805310094,
1059
- "acc_norm": 0.6033057851239669,
1060
- "acc_norm_stderr": 0.044658697805310094
1061
- },
1062
- "harness|hendrycksTest-jurisprudence|5": {
1063
- "acc": 0.5555555555555556,
1064
- "acc_stderr": 0.04803752235190192,
1065
- "acc_norm": 0.5555555555555556,
1066
- "acc_norm_stderr": 0.04803752235190192
1067
- },
1068
- "harness|hendrycksTest-logical_fallacies|5": {
1069
- "acc": 0.5460122699386503,
1070
- "acc_stderr": 0.0391170190467718,
1071
- "acc_norm": 0.5460122699386503,
1072
- "acc_norm_stderr": 0.0391170190467718
1073
- },
1074
- "harness|hendrycksTest-machine_learning|5": {
1075
- "acc": 0.38392857142857145,
1076
- "acc_stderr": 0.04616143075028547,
1077
- "acc_norm": 0.38392857142857145,
1078
- "acc_norm_stderr": 0.04616143075028547
1079
- },
1080
- "harness|hendrycksTest-management|5": {
1081
- "acc": 0.6504854368932039,
1082
- "acc_stderr": 0.047211885060971716,
1083
- "acc_norm": 0.6504854368932039,
1084
- "acc_norm_stderr": 0.047211885060971716
1085
- },
1086
- "harness|hendrycksTest-marketing|5": {
1087
- "acc": 0.7435897435897436,
1088
- "acc_stderr": 0.028605953702004243,
1089
- "acc_norm": 0.7435897435897436,
1090
- "acc_norm_stderr": 0.028605953702004243
1091
- },
1092
- "harness|hendrycksTest-medical_genetics|5": {
1093
- "acc": 0.57,
1094
- "acc_stderr": 0.049756985195624284,
1095
- "acc_norm": 0.57,
1096
- "acc_norm_stderr": 0.049756985195624284
1097
- },
1098
- "harness|hendrycksTest-miscellaneous|5": {
1099
- "acc": 0.6551724137931034,
1100
- "acc_stderr": 0.016997123346113443,
1101
- "acc_norm": 0.6551724137931034,
1102
- "acc_norm_stderr": 0.016997123346113443
1103
- },
1104
- "harness|hendrycksTest-moral_disputes|5": {
1105
- "acc": 0.5346820809248555,
1106
- "acc_stderr": 0.02685425792825887,
1107
- "acc_norm": 0.5346820809248555,
1108
- "acc_norm_stderr": 0.02685425792825887
1109
- },
1110
- "harness|hendrycksTest-moral_scenarios|5": {
1111
- "acc": 0.24804469273743016,
1112
- "acc_stderr": 0.01444415780826144,
1113
- "acc_norm": 0.24804469273743016,
1114
- "acc_norm_stderr": 0.01444415780826144
1115
- },
1116
- "harness|hendrycksTest-nutrition|5": {
1117
- "acc": 0.5784313725490197,
1118
- "acc_stderr": 0.028275490156791462,
1119
- "acc_norm": 0.5784313725490197,
1120
- "acc_norm_stderr": 0.028275490156791462
1121
- },
1122
- "harness|hendrycksTest-philosophy|5": {
1123
- "acc": 0.5401929260450161,
1124
- "acc_stderr": 0.028306190403305696,
1125
- "acc_norm": 0.5401929260450161,
1126
- "acc_norm_stderr": 0.028306190403305696
1127
- },
1128
- "harness|hendrycksTest-prehistory|5": {
1129
- "acc": 0.5462962962962963,
1130
- "acc_stderr": 0.0277012284685426,
1131
- "acc_norm": 0.5462962962962963,
1132
- "acc_norm_stderr": 0.0277012284685426
1133
- },
1134
- "harness|hendrycksTest-professional_accounting|5": {
1135
- "acc": 0.3971631205673759,
1136
- "acc_stderr": 0.029189805673587105,
1137
- "acc_norm": 0.3971631205673759,
1138
- "acc_norm_stderr": 0.029189805673587105
1139
- },
1140
- "harness|hendrycksTest-professional_law|5": {
1141
- "acc": 0.35919165580182527,
1142
- "acc_stderr": 0.012253386187584253,
1143
- "acc_norm": 0.35919165580182527,
1144
- "acc_norm_stderr": 0.012253386187584253
1145
- },
1146
- "harness|hendrycksTest-professional_medicine|5": {
1147
- "acc": 0.4852941176470588,
1148
- "acc_stderr": 0.03035969707904611,
1149
- "acc_norm": 0.4852941176470588,
1150
- "acc_norm_stderr": 0.03035969707904611
1151
- },
1152
- "harness|hendrycksTest-professional_psychology|5": {
1153
- "acc": 0.48366013071895425,
1154
- "acc_stderr": 0.020217030653186457,
1155
- "acc_norm": 0.48366013071895425,
1156
- "acc_norm_stderr": 0.020217030653186457
1157
- },
1158
- "harness|hendrycksTest-public_relations|5": {
1159
- "acc": 0.6,
1160
- "acc_stderr": 0.0469237132203465,
1161
- "acc_norm": 0.6,
1162
- "acc_norm_stderr": 0.0469237132203465
1163
- },
1164
- "harness|hendrycksTest-security_studies|5": {
1165
- "acc": 0.5918367346938775,
1166
- "acc_stderr": 0.03146465712827424,
1167
- "acc_norm": 0.5918367346938775,
1168
- "acc_norm_stderr": 0.03146465712827424
1169
- },
1170
- "harness|hendrycksTest-sociology|5": {
1171
- "acc": 0.6467661691542289,
1172
- "acc_stderr": 0.03379790611796777,
1173
- "acc_norm": 0.6467661691542289,
1174
- "acc_norm_stderr": 0.03379790611796777
1175
- },
1176
- "harness|hendrycksTest-us_foreign_policy|5": {
1177
- "acc": 0.74,
1178
- "acc_stderr": 0.04408440022768079,
1179
- "acc_norm": 0.74,
1180
- "acc_norm_stderr": 0.04408440022768079
1181
- },
1182
- "harness|hendrycksTest-virology|5": {
1183
- "acc": 0.4397590361445783,
1184
- "acc_stderr": 0.03864139923699121,
1185
- "acc_norm": 0.4397590361445783,
1186
- "acc_norm_stderr": 0.03864139923699121
1187
- },
1188
- "harness|hendrycksTest-world_religions|5": {
1189
- "acc": 0.6783625730994152,
1190
- "acc_stderr": 0.03582529442573122,
1191
- "acc_norm": 0.6783625730994152,
1192
- "acc_norm_stderr": 0.03582529442573122
1193
- },
1194
- "harness|truthfulqa:mc|0": {
1195
- "mc1": 0.33659730722154224,
1196
- "mc1_stderr": 0.016542412809494887,
1197
- "mc2": 0.5041278188644791,
1198
- "mc2_stderr": 0.01576728095387478
1199
  }
1200
  }
1201
  ```
 
3
  dataset_summary: "Dataset automatically created during the evaluation run of model\
4
  \ [lmsys/vicuna-7b-v1.5-16k](https://huggingface.co/lmsys/vicuna-7b-v1.5-16k) on\
5
  \ the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard).\n\
6
+ \nThe dataset is composed of 64 configuration, each one coresponding to one of the\
7
+ \ evaluated task.\n\nThe dataset has been created from 2 run(s). Each run can be\
8
  \ found as a specific split in each configuration, the split being named using the\
9
  \ timestamp of the run.The \"train\" split is always pointing to the latest results.\n\
10
  \nAn additional configuration \"results\" store all the aggregated results of the\
 
12
  \ Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)).\n\
13
  \nTo load the details from a run, you can for instance do the following:\n```python\n\
14
  from datasets import load_dataset\ndata = load_dataset(\"open-llm-leaderboard/details_lmsys__vicuna-7b-v1.5-16k\"\
15
+ ,\n\t\"harness_winogrande_5\",\n\tsplit=\"train\")\n```\n\n## Latest results\n\n\
16
+ These are the [latest results from run 2023-10-19T11:48:14.833652](https://huggingface.co/datasets/open-llm-leaderboard/details_lmsys__vicuna-7b-v1.5-16k/blob/main/results_2023-10-19T11-48-14.833652.json)(note\
17
+ \ that their might be results for other tasks in the repos if successive evals didn't\
18
+ \ cover the same tasks. You find each in the results and the \"latest\" split for\
19
+ \ each eval):\n\n```python\n{\n \"all\": {\n \"em\": 0.008389261744966443,\n\
20
+ \ \"em_stderr\": 0.0009340543216867012,\n \"f1\": 0.07163485738255039,\n\
21
+ \ \"f1_stderr\": 0.001675003372186539,\n \"acc\": 0.38778570501079185,\n\
22
+ \ \"acc_stderr\": 0.009750816499753356\n },\n \"harness|drop|3\": {\n\
23
+ \ \"em\": 0.008389261744966443,\n \"em_stderr\": 0.0009340543216867012,\n\
24
+ \ \"f1\": 0.07163485738255039,\n \"f1_stderr\": 0.001675003372186539\n\
25
+ \ },\n \"harness|gsm8k|5\": {\n \"acc\": 0.06444275966641395,\n \
26
+ \ \"acc_stderr\": 0.006763391728488277\n },\n \"harness|winogrande|5\"\
27
+ : {\n \"acc\": 0.7111286503551697,\n \"acc_stderr\": 0.012738241271018436\n\
28
+ \ }\n}\n```"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
29
  repo_url: https://huggingface.co/lmsys/vicuna-7b-v1.5-16k
30
  leaderboard_url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard
31
  point_of_contact: clementine@hf.co
 
38
  - split: latest
39
  path:
40
  - '**/details_harness|arc:challenge|25_2023-08-18T07:58:23.659880.parquet'
41
+ - config_name: harness_drop_3
42
+ data_files:
43
+ - split: 2023_10_19T11_48_14.833652
44
+ path:
45
+ - '**/details_harness|drop|3_2023-10-19T11-48-14.833652.parquet'
46
+ - split: latest
47
+ path:
48
+ - '**/details_harness|drop|3_2023-10-19T11-48-14.833652.parquet'
49
+ - config_name: harness_gsm8k_5
50
+ data_files:
51
+ - split: 2023_10_19T11_48_14.833652
52
+ path:
53
+ - '**/details_harness|gsm8k|5_2023-10-19T11-48-14.833652.parquet'
54
+ - split: latest
55
+ path:
56
+ - '**/details_harness|gsm8k|5_2023-10-19T11-48-14.833652.parquet'
57
  - config_name: harness_hellaswag_10
58
  data_files:
59
  - split: 2023_08_18T07_58_23.659880
 
646
  - split: latest
647
  path:
648
  - '**/details_harness|truthfulqa:mc|0_2023-08-18T07:58:23.659880.parquet'
649
+ - config_name: harness_winogrande_5
650
+ data_files:
651
+ - split: 2023_10_19T11_48_14.833652
652
+ path:
653
+ - '**/details_harness|winogrande|5_2023-10-19T11-48-14.833652.parquet'
654
+ - split: latest
655
+ path:
656
+ - '**/details_harness|winogrande|5_2023-10-19T11-48-14.833652.parquet'
657
  - config_name: results
658
  data_files:
659
  - split: 2023_08_18T07_58_23.659880
660
  path:
661
  - results_2023-08-18T07:58:23.659880.parquet
662
+ - split: 2023_10_19T11_48_14.833652
663
+ path:
664
+ - results_2023-10-19T11-48-14.833652.parquet
665
  - split: latest
666
  path:
667
+ - results_2023-10-19T11-48-14.833652.parquet
668
  ---
669
 
670
  # Dataset Card for Evaluation run of lmsys/vicuna-7b-v1.5-16k
 
681
 
682
  Dataset automatically created during the evaluation run of model [lmsys/vicuna-7b-v1.5-16k](https://huggingface.co/lmsys/vicuna-7b-v1.5-16k) on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard).
683
 
684
+ The dataset is composed of 64 configuration, each one coresponding to one of the evaluated task.
685
 
686
+ The dataset has been created from 2 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
687
 
688
  An additional configuration "results" store all the aggregated results of the run (and is used to compute and display the agregated metrics on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)).
689
 
 
691
  ```python
692
  from datasets import load_dataset
693
  data = load_dataset("open-llm-leaderboard/details_lmsys__vicuna-7b-v1.5-16k",
694
+ "harness_winogrande_5",
695
  split="train")
696
  ```
697
 
698
  ## Latest results
699
 
700
+ These are the [latest results from run 2023-10-19T11:48:14.833652](https://huggingface.co/datasets/open-llm-leaderboard/details_lmsys__vicuna-7b-v1.5-16k/blob/main/results_2023-10-19T11-48-14.833652.json)(note that their might be results for other tasks in the repos if successive evals didn't cover the same tasks. You find each in the results and the "latest" split for each eval):
701
 
702
  ```python
703
  {
704
  "all": {
705
+ "em": 0.008389261744966443,
706
+ "em_stderr": 0.0009340543216867012,
707
+ "f1": 0.07163485738255039,
708
+ "f1_stderr": 0.001675003372186539,
709
+ "acc": 0.38778570501079185,
710
+ "acc_stderr": 0.009750816499753356
711
+ },
712
+ "harness|drop|3": {
713
+ "em": 0.008389261744966443,
714
+ "em_stderr": 0.0009340543216867012,
715
+ "f1": 0.07163485738255039,
716
+ "f1_stderr": 0.001675003372186539
717
+ },
718
+ "harness|gsm8k|5": {
719
+ "acc": 0.06444275966641395,
720
+ "acc_stderr": 0.006763391728488277
721
+ },
722
+ "harness|winogrande|5": {
723
+ "acc": 0.7111286503551697,
724
+ "acc_stderr": 0.012738241271018436
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
725
  }
726
  }
727
  ```