Spaces:
Running
Running
Upload from GitHub Actions: Exclude TruthfulQA from proficiency score
Browse files- evals/backend.py +1 -1
- frontend/src/App.js +219 -90
evals/backend.py
CHANGED
|
@@ -26,7 +26,7 @@ task_metrics = [
|
|
| 26 |
"classification_accuracy",
|
| 27 |
"mmlu_accuracy",
|
| 28 |
"arc_accuracy",
|
| 29 |
-
"truthfulqa_accuracy",
|
| 30 |
"mgsm_accuracy",
|
| 31 |
]
|
| 32 |
|
|
|
|
| 26 |
"classification_accuracy",
|
| 27 |
"mmlu_accuracy",
|
| 28 |
"arc_accuracy",
|
| 29 |
+
# "truthfulqa_accuracy",
|
| 30 |
"mgsm_accuracy",
|
| 31 |
]
|
| 32 |
|
frontend/src/App.js
CHANGED
|
@@ -57,20 +57,47 @@ function App () {
|
|
| 57 |
|
| 58 |
return (
|
| 59 |
<PrimeReactProvider>
|
| 60 |
-
<div
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
|
| 73 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 74 |
GitHub
|
| 75 |
</a>
|
| 76 |
</div>
|
|
@@ -95,30 +122,42 @@ function App () {
|
|
| 95 |
π
|
| 96 |
</span>
|
| 97 |
</div>
|
| 98 |
-
<h1
|
| 99 |
-
|
| 100 |
-
|
| 101 |
-
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
| 105 |
-
|
|
|
|
|
|
|
| 106 |
</h1>
|
| 107 |
-
<p
|
| 108 |
-
|
| 109 |
-
|
| 110 |
-
|
| 111 |
-
|
| 112 |
-
|
| 113 |
-
|
| 114 |
-
|
|
|
|
|
|
|
| 115 |
Comprehensive multilingual evaluation results for AI language models
|
| 116 |
</p>
|
| 117 |
-
|
| 118 |
-
<div
|
| 119 |
-
|
| 120 |
-
|
| 121 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 122 |
onClick={() => setAboutVisible(true)}
|
| 123 |
style={{
|
| 124 |
color: '#666',
|
|
@@ -128,12 +167,12 @@ function App () {
|
|
| 128 |
fontSize: '0.9rem'
|
| 129 |
}}
|
| 130 |
/>
|
| 131 |
-
|
| 132 |
-
<Button
|
| 133 |
-
label=
|
| 134 |
-
className=
|
| 135 |
onClick={() => setContributeVisible(true)}
|
| 136 |
-
tooltip=
|
| 137 |
tooltipOptions={{ position: 'bottom' }}
|
| 138 |
style={{
|
| 139 |
color: '#666',
|
|
@@ -144,7 +183,7 @@ function App () {
|
|
| 144 |
}}
|
| 145 |
/>
|
| 146 |
</div>
|
| 147 |
-
|
| 148 |
{data && (
|
| 149 |
<AutoComplete
|
| 150 |
languages={data?.language_table}
|
|
@@ -164,30 +203,31 @@ function App () {
|
|
| 164 |
>
|
| 165 |
{loading && (
|
| 166 |
<div style={{ width: '100%', textAlign: 'center' }}>
|
| 167 |
-
<i
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 168 |
</div>
|
| 169 |
)}
|
| 170 |
-
{error && <div style={{ width: '100%', textAlign: 'center' }}><p>Error: {error}</p></div>}
|
| 171 |
{data && (
|
| 172 |
<>
|
| 173 |
-
<
|
| 174 |
-
|
| 175 |
-
|
| 176 |
-
|
| 177 |
-
|
| 178 |
-
|
| 179 |
-
|
| 180 |
-
|
| 181 |
-
|
| 182 |
-
|
| 183 |
-
|
| 184 |
-
|
| 185 |
-
totalModels={data.model_table?.length || 0}
|
| 186 |
-
/>
|
| 187 |
-
</div>
|
| 188 |
-
<div style={{ width: '100%' }}>
|
| 189 |
-
<DatasetTable data={data} />
|
| 190 |
-
</div>
|
| 191 |
<div
|
| 192 |
id='figure'
|
| 193 |
style={{
|
|
@@ -196,10 +236,10 @@ function App () {
|
|
| 196 |
}}
|
| 197 |
>
|
| 198 |
<Button
|
| 199 |
-
icon=
|
| 200 |
-
className=
|
| 201 |
onClick={() => setDialogVisible(true)}
|
| 202 |
-
tooltip=
|
| 203 |
style={{
|
| 204 |
position: 'absolute',
|
| 205 |
top: '10px',
|
|
@@ -214,7 +254,7 @@ function App () {
|
|
| 214 |
<LanguagePlot data={data} />,
|
| 215 |
<SpeakerPlot data={data} />,
|
| 216 |
<HistoryPlot data={data} />,
|
| 217 |
-
<CostPlot data={data}
|
| 218 |
]}
|
| 219 |
numScroll={1}
|
| 220 |
numVisible={1}
|
|
@@ -233,35 +273,85 @@ function App () {
|
|
| 233 |
onHide={() => setAboutVisible(false)}
|
| 234 |
style={{ width: '600px' }}
|
| 235 |
modal
|
| 236 |
-
header=
|
| 237 |
>
|
| 238 |
<div>
|
| 239 |
-
<p>
|
|
|
|
|
|
|
|
|
|
| 240 |
<h4>Who is this for?</h4>
|
| 241 |
<ul>
|
| 242 |
-
<li
|
| 243 |
-
|
| 244 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 245 |
</ul>
|
| 246 |
<h4>β‘ Live Updates</h4>
|
| 247 |
-
<p>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 248 |
<h4>Authors</h4>
|
| 249 |
-
<p>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 250 |
<h4>π Links</h4>
|
| 251 |
<p>
|
| 252 |
-
<a
|
| 253 |
-
href=
|
| 254 |
-
target=
|
| 255 |
-
rel=
|
| 256 |
-
style={{
|
| 257 |
-
color: '#666',
|
| 258 |
textDecoration: 'none',
|
| 259 |
display: 'inline-flex',
|
| 260 |
alignItems: 'center',
|
| 261 |
gap: '0.5rem'
|
| 262 |
}}
|
| 263 |
>
|
| 264 |
-
<i className=
|
| 265 |
View source code on GitHub
|
| 266 |
</a>
|
| 267 |
</p>
|
|
@@ -274,16 +364,39 @@ function App () {
|
|
| 274 |
onHide={() => setContributeVisible(false)}
|
| 275 |
style={{ width: '600px' }}
|
| 276 |
modal
|
| 277 |
-
header=
|
| 278 |
>
|
| 279 |
<div>
|
| 280 |
<h4>π Submit Your Model</h4>
|
| 281 |
-
<p>
|
| 282 |
-
|
| 283 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 284 |
<h4>π§ Contribute to Development</h4>
|
| 285 |
-
<p>
|
| 286 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 287 |
</div>
|
| 288 |
</Dialog>
|
| 289 |
|
|
@@ -300,11 +413,27 @@ function App () {
|
|
| 300 |
<div style={{ width: '100%', height: '100%' }}>
|
| 301 |
<Carousel
|
| 302 |
value={[
|
| 303 |
-
<WorldMap
|
| 304 |
-
|
| 305 |
-
|
| 306 |
-
|
| 307 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 308 |
]}
|
| 309 |
numScroll={1}
|
| 310 |
numVisible={1}
|
|
|
|
| 57 |
|
| 58 |
return (
|
| 59 |
<PrimeReactProvider>
|
| 60 |
+
<div
|
| 61 |
+
style={{
|
| 62 |
+
minHeight: '100vh',
|
| 63 |
+
display: 'flex',
|
| 64 |
+
flexDirection: 'column',
|
| 65 |
+
width: '100vw'
|
| 66 |
+
}}
|
| 67 |
+
>
|
| 68 |
+
<div
|
| 69 |
+
style={{
|
| 70 |
+
backgroundColor: '#fff3cd',
|
| 71 |
+
color: '#856404',
|
| 72 |
+
padding: '0.75rem 1.25rem',
|
| 73 |
+
marginBottom: '1rem',
|
| 74 |
+
border: '1px solid #ffeeba',
|
| 75 |
+
borderRadius: '0.25rem',
|
| 76 |
+
textAlign: 'center'
|
| 77 |
+
}}
|
| 78 |
+
>
|
| 79 |
+
<strong>Work in Progress:</strong> This dashboard is currently under
|
| 80 |
+
active development. Evaluation results are not yet final.
|
| 81 |
+
<a
|
| 82 |
+
href='https://github.com/datenlabor-bmz/ai-language-monitor'
|
| 83 |
+
target='_blank'
|
| 84 |
+
rel='noopener noreferrer'
|
| 85 |
+
style={{
|
| 86 |
+
textDecoration: 'none',
|
| 87 |
+
color: '#856404',
|
| 88 |
+
float: 'right',
|
| 89 |
+
fontSize: '1.2rem',
|
| 90 |
+
fontWeight: 'bold',
|
| 91 |
+
padding: '0 0.5rem',
|
| 92 |
+
borderRadius: '3px',
|
| 93 |
+
backgroundColor: 'rgba(255,255,255,0.3)'
|
| 94 |
+
}}
|
| 95 |
+
>
|
| 96 |
+
<i
|
| 97 |
+
className='pi pi-github'
|
| 98 |
+
title='View on GitHub'
|
| 99 |
+
style={{ marginRight: '0.3rem' }}
|
| 100 |
+
/>
|
| 101 |
GitHub
|
| 102 |
</a>
|
| 103 |
</div>
|
|
|
|
| 122 |
π
|
| 123 |
</span>
|
| 124 |
</div>
|
| 125 |
+
<h1
|
| 126 |
+
style={{
|
| 127 |
+
fontSize: '2.5rem',
|
| 128 |
+
fontWeight: '600',
|
| 129 |
+
margin: '1rem 0 0.5rem 0',
|
| 130 |
+
color: '#333',
|
| 131 |
+
letterSpacing: '-0.01em'
|
| 132 |
+
}}
|
| 133 |
+
>
|
| 134 |
+
AI Language Proficiency Monitor
|
| 135 |
</h1>
|
| 136 |
+
<p
|
| 137 |
+
style={{
|
| 138 |
+
fontSize: '1.1rem',
|
| 139 |
+
color: '#666',
|
| 140 |
+
margin: '0 0 2.5rem 0',
|
| 141 |
+
fontWeight: '400',
|
| 142 |
+
maxWidth: '700px',
|
| 143 |
+
lineHeight: '1.5'
|
| 144 |
+
}}
|
| 145 |
+
>
|
| 146 |
Comprehensive multilingual evaluation results for AI language models
|
| 147 |
</p>
|
| 148 |
+
|
| 149 |
+
<div
|
| 150 |
+
style={{
|
| 151 |
+
display: 'flex',
|
| 152 |
+
gap: '1rem',
|
| 153 |
+
marginBottom: '1.5rem',
|
| 154 |
+
flexWrap: 'wrap',
|
| 155 |
+
justifyContent: 'center'
|
| 156 |
+
}}
|
| 157 |
+
>
|
| 158 |
+
<Button
|
| 159 |
+
label='π About this tool'
|
| 160 |
+
className='p-button-text'
|
| 161 |
onClick={() => setAboutVisible(true)}
|
| 162 |
style={{
|
| 163 |
color: '#666',
|
|
|
|
| 167 |
fontSize: '0.9rem'
|
| 168 |
}}
|
| 169 |
/>
|
| 170 |
+
|
| 171 |
+
<Button
|
| 172 |
+
label='π Add your model (soon)'
|
| 173 |
+
className='p-button-text'
|
| 174 |
onClick={() => setContributeVisible(true)}
|
| 175 |
+
tooltip='This feature is on our roadmap and will be available soon.'
|
| 176 |
tooltipOptions={{ position: 'bottom' }}
|
| 177 |
style={{
|
| 178 |
color: '#666',
|
|
|
|
| 183 |
}}
|
| 184 |
/>
|
| 185 |
</div>
|
| 186 |
+
|
| 187 |
{data && (
|
| 188 |
<AutoComplete
|
| 189 |
languages={data?.language_table}
|
|
|
|
| 203 |
>
|
| 204 |
{loading && (
|
| 205 |
<div style={{ width: '100%', textAlign: 'center' }}>
|
| 206 |
+
<i
|
| 207 |
+
className='pi pi-spinner pi-spin'
|
| 208 |
+
style={{ fontSize: '4rem' }}
|
| 209 |
+
/>
|
| 210 |
+
</div>
|
| 211 |
+
)}
|
| 212 |
+
{error && (
|
| 213 |
+
<div style={{ width: '100%', textAlign: 'center' }}>
|
| 214 |
+
<p>Error: {error}</p>
|
| 215 |
</div>
|
| 216 |
)}
|
|
|
|
| 217 |
{data && (
|
| 218 |
<>
|
| 219 |
+
<ModelTable
|
| 220 |
+
data={data.model_table}
|
| 221 |
+
selectedLanguages={selectedLanguages}
|
| 222 |
+
allLanguages={data.language_table || []}
|
| 223 |
+
/>
|
| 224 |
+
<LanguageTable
|
| 225 |
+
data={data.language_table}
|
| 226 |
+
selectedLanguages={selectedLanguages}
|
| 227 |
+
setSelectedLanguages={setSelectedLanguages}
|
| 228 |
+
totalModels={data.model_table?.length || 0}
|
| 229 |
+
/>
|
| 230 |
+
<DatasetTable data={data} />
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 231 |
<div
|
| 232 |
id='figure'
|
| 233 |
style={{
|
|
|
|
| 236 |
}}
|
| 237 |
>
|
| 238 |
<Button
|
| 239 |
+
icon='pi pi-external-link'
|
| 240 |
+
className='p-button-text p-button-plain'
|
| 241 |
onClick={() => setDialogVisible(true)}
|
| 242 |
+
tooltip='Open in larger view'
|
| 243 |
style={{
|
| 244 |
position: 'absolute',
|
| 245 |
top: '10px',
|
|
|
|
| 254 |
<LanguagePlot data={data} />,
|
| 255 |
<SpeakerPlot data={data} />,
|
| 256 |
<HistoryPlot data={data} />,
|
| 257 |
+
<CostPlot data={data} />
|
| 258 |
]}
|
| 259 |
numScroll={1}
|
| 260 |
numVisible={1}
|
|
|
|
| 273 |
onHide={() => setAboutVisible(false)}
|
| 274 |
style={{ width: '600px' }}
|
| 275 |
modal
|
| 276 |
+
header='About this tool'
|
| 277 |
>
|
| 278 |
<div>
|
| 279 |
+
<p>
|
| 280 |
+
The <i>AI Language Proficiency Monitor</i> presents comprehensive
|
| 281 |
+
multilingual evaluation results of AI language models.
|
| 282 |
+
</p>
|
| 283 |
<h4>Who is this for?</h4>
|
| 284 |
<ul>
|
| 285 |
+
<li>
|
| 286 |
+
<b>Practitioners</b> can pick the best model for a given
|
| 287 |
+
language.
|
| 288 |
+
</li>
|
| 289 |
+
<li>
|
| 290 |
+
<b>Policymakers and funders</b> can identify and prioritize
|
| 291 |
+
neglected languages.
|
| 292 |
+
</li>
|
| 293 |
+
<li>
|
| 294 |
+
<b>Model developers</b> can compete on our{' '}
|
| 295 |
+
<i>AI Language Proficiency</i> metric.
|
| 296 |
+
</li>
|
| 297 |
</ul>
|
| 298 |
<h4>β‘ Live Updates</h4>
|
| 299 |
+
<p>
|
| 300 |
+
Benchmark results automatically refresh every night and include
|
| 301 |
+
the most popular models from{' '}
|
| 302 |
+
<a
|
| 303 |
+
href='https://openrouter.ai'
|
| 304 |
+
target='_blank'
|
| 305 |
+
rel='noopener noreferrer'
|
| 306 |
+
>
|
| 307 |
+
OpenRouter
|
| 308 |
+
</a>
|
| 309 |
+
, plus community-submitted models.
|
| 310 |
+
</p>
|
| 311 |
<h4>Authors</h4>
|
| 312 |
+
<p>
|
| 313 |
+
The AI Language Proficiency Monitor is a collaboration between
|
| 314 |
+
BMZ's{' '}
|
| 315 |
+
<a
|
| 316 |
+
href='https://www.bmz-digital.global/en/overview-of-initiatives/the-bmz-data-lab/'
|
| 317 |
+
target='_blank'
|
| 318 |
+
rel='noopener noreferrer'
|
| 319 |
+
>
|
| 320 |
+
Data Lab
|
| 321 |
+
</a>
|
| 322 |
+
, the BMZ-Initiative{' '}
|
| 323 |
+
<a
|
| 324 |
+
href='https://www.bmz-digital.global/en/overview-of-initiatives/fair-forward/'
|
| 325 |
+
target='_blank'
|
| 326 |
+
rel='noopener noreferrer'
|
| 327 |
+
>
|
| 328 |
+
Fair Forward
|
| 329 |
+
</a>{' '}
|
| 330 |
+
(implemented by GIZ), and the{' '}
|
| 331 |
+
<a
|
| 332 |
+
href='https://www.dfki.de/en/web/research/research-departments/multilinguality-and-language-technology/ee-team'
|
| 333 |
+
target='_blank'
|
| 334 |
+
rel='noopener noreferrer'
|
| 335 |
+
>
|
| 336 |
+
E&E group
|
| 337 |
+
</a>{' '}
|
| 338 |
+
of DFKI's Multilinguality and Language Technology Lab.
|
| 339 |
+
</p>
|
| 340 |
<h4>π Links</h4>
|
| 341 |
<p>
|
| 342 |
+
<a
|
| 343 |
+
href='https://github.com/datenlabor-bmz/ai-language-monitor'
|
| 344 |
+
target='_blank'
|
| 345 |
+
rel='noopener noreferrer'
|
| 346 |
+
style={{
|
| 347 |
+
color: '#666',
|
| 348 |
textDecoration: 'none',
|
| 349 |
display: 'inline-flex',
|
| 350 |
alignItems: 'center',
|
| 351 |
gap: '0.5rem'
|
| 352 |
}}
|
| 353 |
>
|
| 354 |
+
<i className='pi pi-github' style={{ fontSize: '1.2rem' }} />
|
| 355 |
View source code on GitHub
|
| 356 |
</a>
|
| 357 |
</p>
|
|
|
|
| 364 |
onHide={() => setContributeVisible(false)}
|
| 365 |
style={{ width: '600px' }}
|
| 366 |
modal
|
| 367 |
+
header='Add your model & Contribute'
|
| 368 |
>
|
| 369 |
<div>
|
| 370 |
<h4>π Submit Your Model</h4>
|
| 371 |
+
<p>
|
| 372 |
+
Have a custom fine-tuned model you'd like to see on the
|
| 373 |
+
leaderboard?
|
| 374 |
+
</p>
|
| 375 |
+
<p>
|
| 376 |
+
<a
|
| 377 |
+
href='https://forms.gle/ckvY9pS7XLcHYnaV8'
|
| 378 |
+
target='_blank'
|
| 379 |
+
rel='noopener noreferrer'
|
| 380 |
+
style={{ color: '#28a745', fontWeight: 'bold' }}
|
| 381 |
+
>
|
| 382 |
+
β Submit your model here
|
| 383 |
+
</a>
|
| 384 |
+
</p>
|
| 385 |
+
|
| 386 |
<h4>π§ Contribute to Development</h4>
|
| 387 |
+
<p>
|
| 388 |
+
Help us expand language coverage and add new evaluation tasks:
|
| 389 |
+
</p>
|
| 390 |
+
<p>
|
| 391 |
+
<a
|
| 392 |
+
href='https://github.com/datenlabor-bmz/ai-language-monitor/blob/main/CONTRIBUTING.md'
|
| 393 |
+
target='_blank'
|
| 394 |
+
rel='noopener noreferrer'
|
| 395 |
+
style={{ color: '#007bff', fontWeight: 'bold' }}
|
| 396 |
+
>
|
| 397 |
+
β Contribution guidelines
|
| 398 |
+
</a>
|
| 399 |
+
</p>
|
| 400 |
</div>
|
| 401 |
</Dialog>
|
| 402 |
|
|
|
|
| 413 |
<div style={{ width: '100%', height: '100%' }}>
|
| 414 |
<Carousel
|
| 415 |
value={[
|
| 416 |
+
<WorldMap
|
| 417 |
+
data={data.countries}
|
| 418 |
+
width={windowWidth * 0.7}
|
| 419 |
+
height={windowHeight * 0.6}
|
| 420 |
+
/>,
|
| 421 |
+
<LanguagePlot
|
| 422 |
+
data={data}
|
| 423 |
+
width={windowWidth * 0.7}
|
| 424 |
+
height={windowHeight * 0.6}
|
| 425 |
+
/>,
|
| 426 |
+
<SpeakerPlot
|
| 427 |
+
data={data}
|
| 428 |
+
width={windowWidth * 0.7}
|
| 429 |
+
height={windowHeight * 0.6}
|
| 430 |
+
/>,
|
| 431 |
+
<HistoryPlot
|
| 432 |
+
data={data}
|
| 433 |
+
width={windowWidth * 0.7}
|
| 434 |
+
height={windowHeight * 0.6}
|
| 435 |
+
/>,
|
| 436 |
+
<CostPlot data={data} />
|
| 437 |
]}
|
| 438 |
numScroll={1}
|
| 439 |
numVisible={1}
|