melaseddik
commited on
Commit
•
e29972d
1
Parent(s):
bad033a
Update README.md
Browse files
README.md
CHANGED
@@ -167,7 +167,7 @@ print(tokenizer.decode(outputs[0]))
|
|
167 |
</table>
|
168 |
|
169 |
|
170 |
-
<table border="1" style="width: 100%; text-align: center;">
|
171 |
<colgroup>
|
172 |
<col style="width: 10%;">
|
173 |
<col style="width: 10%;">
|
@@ -229,11 +229,124 @@ print(tokenizer.decode(outputs[0]))
|
|
229 |
<td>16.1</td>
|
230 |
<td>36.4</td>
|
231 |
</tr>
|
232 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
233 |
</tbody>
|
234 |
</table>
|
235 |
|
236 |
|
|
|
237 |
# Citation
|
238 |
|
239 |
|
|
|
167 |
</table>
|
168 |
|
169 |
|
170 |
+
<table border="1" style="width: 100%; text-align: center; border-collapse: collapse;">
|
171 |
<colgroup>
|
172 |
<col style="width: 10%;">
|
173 |
<col style="width: 10%;">
|
|
|
229 |
<td>16.1</td>
|
230 |
<td>36.4</td>
|
231 |
</tr>
|
232 |
+
<tr>
|
233 |
+
<td rowspan="2">Math</td>
|
234 |
+
<td>GSM8K (5-shot)</td>
|
235 |
+
<td>49.4</td>
|
236 |
+
<td>77.9</td>
|
237 |
+
<td>82.9</td>
|
238 |
+
<td>76.2</td>
|
239 |
+
<td>69.1</td>
|
240 |
+
<td>63.8</td>
|
241 |
+
<td>55.3</td>
|
242 |
+
<td>81.4</td>
|
243 |
+
</tr>
|
244 |
+
<tr>
|
245 |
+
<td>MATH(4-shot)</td>
|
246 |
+
<td>4.1</td>
|
247 |
+
<td>17.5</td>
|
248 |
+
<td>15.5</td>
|
249 |
+
<td>18.0</td>
|
250 |
+
<td>0</td>
|
251 |
+
<td>9.2</td>
|
252 |
+
<td>4.9</td>
|
253 |
+
<td>22.9</td>
|
254 |
+
</tr>
|
255 |
+
<tr>
|
256 |
+
<td rowspan="4">Reasoning</td>
|
257 |
+
<td>Arc Challenge (25-shot)</td>
|
258 |
+
<td>53.4</td>
|
259 |
+
<td>57.4</td>
|
260 |
+
<td>59.0</td>
|
261 |
+
<td>59.6</td>
|
262 |
+
<td>63.7</td>
|
263 |
+
<td>58.2</td>
|
264 |
+
<td>60.6</td>
|
265 |
+
<td>62.6</td>
|
266 |
+
</tr>
|
267 |
+
<tr>
|
268 |
+
<td>GPQA (0-shot)</td>
|
269 |
+
<td>31.0</td>
|
270 |
+
<td>31.9</td>
|
271 |
+
<td>33.0</td>
|
272 |
+
<td>35.5</td>
|
273 |
+
<td>0</td>
|
274 |
+
<td>36.6</td>
|
275 |
+
<td>28.8</td>
|
276 |
+
<td>34.1</td>
|
277 |
+
</tr>
|
278 |
+
<tr>
|
279 |
+
<td>MUSR (0-shot)</td>
|
280 |
+
<td>38.0</td>
|
281 |
+
<td>44.1</td>
|
282 |
+
<td>44.2</td>
|
283 |
+
<td>47.3</td>
|
284 |
+
<td>0</td>
|
285 |
+
<td>43.3</td>
|
286 |
+
<td>39.2</td>
|
287 |
+
<td>44.2</td>
|
288 |
+
</tr>
|
289 |
+
<tr>
|
290 |
+
<td>BBH (3-shot)</td>
|
291 |
+
<td>46.5</td>
|
292 |
+
<td>53.3</td>
|
293 |
+
<td>54.0</td>
|
294 |
+
<td>51.0</td>
|
295 |
+
<td>0</td>
|
296 |
+
<td>51.3</td>
|
297 |
+
<td>50.2</td>
|
298 |
+
<td>59.7</td>
|
299 |
+
</tr>
|
300 |
+
<tr>
|
301 |
+
<td rowspan="4">CommonSense Understanding</td>
|
302 |
+
<td>PIQA (0-shot)</td>
|
303 |
+
<td>80.3</td>
|
304 |
+
<td>79.8</td>
|
305 |
+
<td>78.7</td>
|
306 |
+
<td>77.7</td>
|
307 |
+
<td>81.4</td>
|
308 |
+
<td>79.8</td>
|
309 |
+
<td>81.4</td>
|
310 |
+
<td>79.1</td>
|
311 |
+
</tr>
|
312 |
+
<tr>
|
313 |
+
<td>SciQ (0-shot)</td>
|
314 |
+
<td>96.3</td>
|
315 |
+
<td>95.9</td>
|
316 |
+
<td>96.6</td>
|
317 |
+
<td>95.3</td>
|
318 |
+
<td>97.2</td>
|
319 |
+
<td>95.8</td>
|
320 |
+
<td>96.4</td>
|
321 |
+
<td>96.0</td>
|
322 |
+
</tr>
|
323 |
+
<tr>
|
324 |
+
<td>Winogrande (0-shot)</td>
|
325 |
+
<td>74.0</td>
|
326 |
+
<td>72.1</td>
|
327 |
+
<td>72.9</td>
|
328 |
+
<td>71.0</td>
|
329 |
+
<td>74.2</td>
|
330 |
+
<td>72.7</td>
|
331 |
+
<td>73.2</td>
|
332 |
+
<td>73.6</td>
|
333 |
+
</tr>
|
334 |
+
<tr>
|
335 |
+
<td>OpenbookQA (0-shot)</td>
|
336 |
+
<td>33.4</td>
|
337 |
+
<td>35.2</td>
|
338 |
+
<td>33.6</td>
|
339 |
+
<td>31.4</td>
|
340 |
+
<td>34.0</td>
|
341 |
+
<td>35.4</td>
|
342 |
+
<td>36.4</td>
|
343 |
+
<td>34.0</td>
|
344 |
+
</tr>
|
345 |
</tbody>
|
346 |
</table>
|
347 |
|
348 |
|
349 |
+
|
350 |
# Citation
|
351 |
|
352 |
|