melaseddik
commited on
Commit
•
a1cf49e
1
Parent(s):
06807eb
Update README.md
Browse files
README.md
CHANGED
@@ -126,10 +126,6 @@ print(tokenizer.decode(outputs[0]))
|
|
126 |
<col style="width: 7%;">
|
127 |
<col style="width: 7%;">
|
128 |
<col style="background-color: rgba(80, 15, 213, 0.5); width: 7%;">
|
129 |
-
<col style="width: 7%;">
|
130 |
-
<col style="width: 7%;">
|
131 |
-
<col style="width: 7%;">
|
132 |
-
<col style="background-color: rgba(80, 15, 213, 0.5); width: 7%;">
|
133 |
</colgroup>
|
134 |
<thead>
|
135 |
<tr>
|
@@ -139,10 +135,6 @@ print(tokenizer.decode(outputs[0]))
|
|
139 |
<th>Qwen2-7B</th>
|
140 |
<th>Qwen2.5-7B</th>
|
141 |
<th>Falcon3-7B-Base</th>
|
142 |
-
<th>Gemma2-9B</th>
|
143 |
-
<th>Yi1.5-9B</th>
|
144 |
-
<th>Mistral-NeMo-12B</th>
|
145 |
-
<th>Falcon3-10B-Base</th>
|
146 |
</tr>
|
147 |
</thead>
|
148 |
<tbody>
|
@@ -153,10 +145,6 @@ print(tokenizer.decode(outputs[0]))
|
|
153 |
<td>70.4</td>
|
154 |
<td>74.2</td>
|
155 |
<td>67.5</td>
|
156 |
-
<td>0</td>
|
157 |
-
<td>69.6</td>
|
158 |
-
<td>68.8</td>
|
159 |
-
<td>73.1</td>
|
160 |
</tr>
|
161 |
<tr>
|
162 |
<td>MMLU-PRO (5-shot)</td>
|
@@ -164,10 +152,6 @@ print(tokenizer.decode(outputs[0]))
|
|
164 |
<td>42.1</td>
|
165 |
<td>43.5</td>
|
166 |
<td>39.2</td>
|
167 |
-
<td>0</td>
|
168 |
-
<td>39.3</td>
|
169 |
-
<td>34.7</td>
|
170 |
-
<td>42.5</td>
|
171 |
</tr>
|
172 |
<tr>
|
173 |
<td>IFEval</td>
|
@@ -175,10 +159,6 @@ print(tokenizer.decode(outputs[0]))
|
|
175 |
<td>30.6</td>
|
176 |
<td>33.9</td>
|
177 |
<td>34.3</td>
|
178 |
-
<td>0</td>
|
179 |
-
<td>29.1</td>
|
180 |
-
<td>16.1</td>
|
181 |
-
<td>36.4</td>
|
182 |
</tr>
|
183 |
<tr>
|
184 |
<td rowspan="2">Math</td>
|
@@ -187,10 +167,6 @@ print(tokenizer.decode(outputs[0]))
|
|
187 |
<td>77.9</td>
|
188 |
<td>82.9</td>
|
189 |
<td>76.2</td>
|
190 |
-
<td>69.1</td>
|
191 |
-
<td>63.8</td>
|
192 |
-
<td>55.3</td>
|
193 |
-
<td>81.4</td>
|
194 |
</tr>
|
195 |
<tr>
|
196 |
<td>MATH(4-shot)</td>
|
@@ -198,10 +174,6 @@ print(tokenizer.decode(outputs[0]))
|
|
198 |
<td>17.5</td>
|
199 |
<td>15.5</td>
|
200 |
<td>18.0</td>
|
201 |
-
<td>0</td>
|
202 |
-
<td>9.2</td>
|
203 |
-
<td>4.9</td>
|
204 |
-
<td>22.9</td>
|
205 |
</tr>
|
206 |
<tr>
|
207 |
<td rowspan="4">Reasoning</td>
|
@@ -210,10 +182,6 @@ print(tokenizer.decode(outputs[0]))
|
|
210 |
<td>57.4</td>
|
211 |
<td>59.0</td>
|
212 |
<td>59.6</td>
|
213 |
-
<td>63.7</td>
|
214 |
-
<td>58.2</td>
|
215 |
-
<td>60.6</td>
|
216 |
-
<td>62.6</td>
|
217 |
</tr>
|
218 |
<tr>
|
219 |
<td>GPQA (0-shot)</td>
|
@@ -221,10 +189,6 @@ print(tokenizer.decode(outputs[0]))
|
|
221 |
<td>31.9</td>
|
222 |
<td>33.0</td>
|
223 |
<td>35.5</td>
|
224 |
-
<td>0</td>
|
225 |
-
<td>36.6</td>
|
226 |
-
<td>28.8</td>
|
227 |
-
<td>34.1</td>
|
228 |
</tr>
|
229 |
<tr>
|
230 |
<td>MUSR (0-shot)</td>
|
@@ -232,10 +196,6 @@ print(tokenizer.decode(outputs[0]))
|
|
232 |
<td>44.1</td>
|
233 |
<td>44.2</td>
|
234 |
<td>47.3</td>
|
235 |
-
<td>0</td>
|
236 |
-
<td>43.3</td>
|
237 |
-
<td>39.2</td>
|
238 |
-
<td>44.2</td>
|
239 |
</tr>
|
240 |
<tr>
|
241 |
<td>BBH (3-shot)</td>
|
@@ -243,10 +203,6 @@ print(tokenizer.decode(outputs[0]))
|
|
243 |
<td>53.3</td>
|
244 |
<td>54.0</td>
|
245 |
<td>51.0</td>
|
246 |
-
<td>0</td>
|
247 |
-
<td>51.3</td>
|
248 |
-
<td>50.2</td>
|
249 |
-
<td>59.7</td>
|
250 |
</tr>
|
251 |
<tr>
|
252 |
<td rowspan="4">CommonSense Understanding</td>
|
@@ -255,10 +211,6 @@ print(tokenizer.decode(outputs[0]))
|
|
255 |
<td>79.8</td>
|
256 |
<td>78.7</td>
|
257 |
<td>77.7</td>
|
258 |
-
<td>81.4</td>
|
259 |
-
<td>79.8</td>
|
260 |
-
<td>81.4</td>
|
261 |
-
<td>79.1</td>
|
262 |
</tr>
|
263 |
<tr>
|
264 |
<td>SciQ (0-shot)</td>
|
@@ -266,10 +218,6 @@ print(tokenizer.decode(outputs[0]))
|
|
266 |
<td>95.9</td>
|
267 |
<td>96.6</td>
|
268 |
<td>95.3</td>
|
269 |
-
<td>97.2</td>
|
270 |
-
<td>95.8</td>
|
271 |
-
<td>96.4</td>
|
272 |
-
<td>96.0</td>
|
273 |
</tr>
|
274 |
<tr>
|
275 |
<td>Winogrande (0-shot)</td>
|
@@ -277,10 +225,6 @@ print(tokenizer.decode(outputs[0]))
|
|
277 |
<td>72.1</td>
|
278 |
<td>72.9</td>
|
279 |
<td>71.0</td>
|
280 |
-
<td>74.2</td>
|
281 |
-
<td>72.7</td>
|
282 |
-
<td>73.2</td>
|
283 |
-
<td>73.6</td>
|
284 |
</tr>
|
285 |
<tr>
|
286 |
<td>OpenbookQA (0-shot)</td>
|
@@ -288,10 +232,6 @@ print(tokenizer.decode(outputs[0]))
|
|
288 |
<td>35.2</td>
|
289 |
<td>33.6</td>
|
290 |
<td>31.4</td>
|
291 |
-
<td>34.0</td>
|
292 |
-
<td>35.4</td>
|
293 |
-
<td>36.4</td>
|
294 |
-
<td>34.0</td>
|
295 |
</tr>
|
296 |
</tbody>
|
297 |
</table>
|
|
|
126 |
<col style="width: 7%;">
|
127 |
<col style="width: 7%;">
|
128 |
<col style="background-color: rgba(80, 15, 213, 0.5); width: 7%;">
|
|
|
|
|
|
|
|
|
129 |
</colgroup>
|
130 |
<thead>
|
131 |
<tr>
|
|
|
135 |
<th>Qwen2-7B</th>
|
136 |
<th>Qwen2.5-7B</th>
|
137 |
<th>Falcon3-7B-Base</th>
|
|
|
|
|
|
|
|
|
138 |
</tr>
|
139 |
</thead>
|
140 |
<tbody>
|
|
|
145 |
<td>70.4</td>
|
146 |
<td>74.2</td>
|
147 |
<td>67.5</td>
|
|
|
|
|
|
|
|
|
148 |
</tr>
|
149 |
<tr>
|
150 |
<td>MMLU-PRO (5-shot)</td>
|
|
|
152 |
<td>42.1</td>
|
153 |
<td>43.5</td>
|
154 |
<td>39.2</td>
|
|
|
|
|
|
|
|
|
155 |
</tr>
|
156 |
<tr>
|
157 |
<td>IFEval</td>
|
|
|
159 |
<td>30.6</td>
|
160 |
<td>33.9</td>
|
161 |
<td>34.3</td>
|
|
|
|
|
|
|
|
|
162 |
</tr>
|
163 |
<tr>
|
164 |
<td rowspan="2">Math</td>
|
|
|
167 |
<td>77.9</td>
|
168 |
<td>82.9</td>
|
169 |
<td>76.2</td>
|
|
|
|
|
|
|
|
|
170 |
</tr>
|
171 |
<tr>
|
172 |
<td>MATH(4-shot)</td>
|
|
|
174 |
<td>17.5</td>
|
175 |
<td>15.5</td>
|
176 |
<td>18.0</td>
|
|
|
|
|
|
|
|
|
177 |
</tr>
|
178 |
<tr>
|
179 |
<td rowspan="4">Reasoning</td>
|
|
|
182 |
<td>57.4</td>
|
183 |
<td>59.0</td>
|
184 |
<td>59.6</td>
|
|
|
|
|
|
|
|
|
185 |
</tr>
|
186 |
<tr>
|
187 |
<td>GPQA (0-shot)</td>
|
|
|
189 |
<td>31.9</td>
|
190 |
<td>33.0</td>
|
191 |
<td>35.5</td>
|
|
|
|
|
|
|
|
|
192 |
</tr>
|
193 |
<tr>
|
194 |
<td>MUSR (0-shot)</td>
|
|
|
196 |
<td>44.1</td>
|
197 |
<td>44.2</td>
|
198 |
<td>47.3</td>
|
|
|
|
|
|
|
|
|
199 |
</tr>
|
200 |
<tr>
|
201 |
<td>BBH (3-shot)</td>
|
|
|
203 |
<td>53.3</td>
|
204 |
<td>54.0</td>
|
205 |
<td>51.0</td>
|
|
|
|
|
|
|
|
|
206 |
</tr>
|
207 |
<tr>
|
208 |
<td rowspan="4">CommonSense Understanding</td>
|
|
|
211 |
<td>79.8</td>
|
212 |
<td>78.7</td>
|
213 |
<td>77.7</td>
|
|
|
|
|
|
|
|
|
214 |
</tr>
|
215 |
<tr>
|
216 |
<td>SciQ (0-shot)</td>
|
|
|
218 |
<td>95.9</td>
|
219 |
<td>96.6</td>
|
220 |
<td>95.3</td>
|
|
|
|
|
|
|
|
|
221 |
</tr>
|
222 |
<tr>
|
223 |
<td>Winogrande (0-shot)</td>
|
|
|
225 |
<td>72.1</td>
|
226 |
<td>72.9</td>
|
227 |
<td>71.0</td>
|
|
|
|
|
|
|
|
|
228 |
</tr>
|
229 |
<tr>
|
230 |
<td>OpenbookQA (0-shot)</td>
|
|
|
232 |
<td>35.2</td>
|
233 |
<td>33.6</td>
|
234 |
<td>31.4</td>
|
|
|
|
|
|
|
|
|
235 |
</tr>
|
236 |
</tbody>
|
237 |
</table>
|