Update README.md
Browse files
README.md
CHANGED
|
@@ -166,4 +166,123 @@ The model was evaluated on the OpenLLMv1 leaderboard task, using [lm-evaluation-
|
|
| 166 |
```
|
| 167 |
|
| 168 |
|
| 169 |
-
</details>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 166 |
```
|
| 167 |
|
| 168 |
|
| 169 |
+
</details>
|
| 170 |
+
|
| 171 |
+
|
| 172 |
+
|
| 173 |
+
|
| 174 |
+
|
| 175 |
+
|
| 176 |
+
|
| 177 |
+
|
| 178 |
+
### Accuracy
|
| 179 |
+
|
| 180 |
+
<table>
|
| 181 |
+
<thead>
|
| 182 |
+
<tr>
|
| 183 |
+
<th>Category</th>
|
| 184 |
+
<th>Metric</th>
|
| 185 |
+
<th>meta-llama/Llama-3.1-70B-Instruct</th>
|
| 186 |
+
<th>nm-testing/Llama-3.1-70B-Instruct-FP8-block</th>
|
| 187 |
+
<th>Recovery (%)</th>
|
| 188 |
+
</tr>
|
| 189 |
+
</thead>
|
| 190 |
+
<tbody>
|
| 191 |
+
<!-- OpenLLM Leaderboard V1 -->
|
| 192 |
+
<tr>
|
| 193 |
+
<td rowspan="7"><b>OpenLLM V1</b></td>
|
| 194 |
+
<td>ARC-Challenge (Acc-Norm, 25-shot)</td>
|
| 195 |
+
<td>abc</td>
|
| 196 |
+
<td>ijk</td>
|
| 197 |
+
<td>xyz</td>
|
| 198 |
+
</tr>
|
| 199 |
+
<tr>
|
| 200 |
+
<td>GSM8K (Strict-Match, 5-shot)</td>
|
| 201 |
+
<td>abc</td>
|
| 202 |
+
<td>ijk</td>
|
| 203 |
+
<td>xyz</td>
|
| 204 |
+
</tr>
|
| 205 |
+
<tr>
|
| 206 |
+
<td>HellaSwag (Acc-Norm, 10-shot)</td>
|
| 207 |
+
<td>abc</td>
|
| 208 |
+
<td>ijk</td>
|
| 209 |
+
<td>xyz</td>
|
| 210 |
+
</tr>
|
| 211 |
+
<tr>
|
| 212 |
+
<td>MMLU (Acc, 5-shot)</td>
|
| 213 |
+
<td>abc</td>
|
| 214 |
+
<td>ijk</td>
|
| 215 |
+
<td>xyz</td>
|
| 216 |
+
</tr>
|
| 217 |
+
<tr>
|
| 218 |
+
<td>TruthfulQA (MC2, 0-shot)</td>
|
| 219 |
+
<td>abc</td>
|
| 220 |
+
<td>ijk</td>
|
| 221 |
+
<td>xyz</td>
|
| 222 |
+
</tr>
|
| 223 |
+
<tr>
|
| 224 |
+
<td>Winogrande (Acc, 5-shot)</td>
|
| 225 |
+
<td>abc</td>
|
| 226 |
+
<td>ijk</td>
|
| 227 |
+
<td>xyz</td>
|
| 228 |
+
</tr>
|
| 229 |
+
<tr>
|
| 230 |
+
<td><b>Average Score</b></td>
|
| 231 |
+
<td><b>abc</b></td>
|
| 232 |
+
<td><b>ijk</b></td>
|
| 233 |
+
<td><b>xyz</b></td>
|
| 234 |
+
</tr>
|
| 235 |
+
<!-- OpenLLM Leaderboard V2 -->
|
| 236 |
+
<tr>
|
| 237 |
+
<td rowspan="7"><b>OpenLLM V2</b></td>
|
| 238 |
+
<td>IFEval (Inst Level Strict Acc, 0-shot)</td>
|
| 239 |
+
<td>abc</td>
|
| 240 |
+
<td>ijk</td>
|
| 241 |
+
<td>xyz</td>
|
| 242 |
+
</tr>
|
| 243 |
+
<tr>
|
| 244 |
+
<td>BBH (Acc-Norm, 3-shot)</td>
|
| 245 |
+
<td>abc</td>
|
| 246 |
+
<td>ijk</td>
|
| 247 |
+
<td>xyz</td>
|
| 248 |
+
</tr>
|
| 249 |
+
<tr>
|
| 250 |
+
<td>Math-Hard (Exact-Match, 4-shot)</td>
|
| 251 |
+
<td>abc</td>
|
| 252 |
+
<td>ijk</td>
|
| 253 |
+
<td>xyz</td>
|
| 254 |
+
</tr>
|
| 255 |
+
<tr>
|
| 256 |
+
<td>GPQA (Acc-Norm, 0-shot)</td>
|
| 257 |
+
<td>abc</td>
|
| 258 |
+
<td>ijk</td>
|
| 259 |
+
<td>xyz</td>
|
| 260 |
+
</tr>
|
| 261 |
+
<tr>
|
| 262 |
+
<td>MUSR (Acc-Norm, 0-shot)</td>
|
| 263 |
+
<td>abc</td>
|
| 264 |
+
<td>ijk</td>
|
| 265 |
+
<td>xyz</td>
|
| 266 |
+
</tr>
|
| 267 |
+
<tr>
|
| 268 |
+
<td>MMLU-Pro (Acc, 5-shot)</td>
|
| 269 |
+
<td>abc</td>
|
| 270 |
+
<td>ijk</td>
|
| 271 |
+
<td>xyz</td>
|
| 272 |
+
</tr>
|
| 273 |
+
<tr>
|
| 274 |
+
<td><b>Average Score</b></td>
|
| 275 |
+
<td><b>abc</b></td>
|
| 276 |
+
<td><b>ijk</b></td>
|
| 277 |
+
<td><b>xyz</b></td>
|
| 278 |
+
</tr>
|
| 279 |
+
<!-- HumanEval -->
|
| 280 |
+
<tr>
|
| 281 |
+
<td rowspan="2"><b>Coding</b></td>
|
| 282 |
+
<td>HumanEval Pass@1</td>
|
| 283 |
+
<td>abc</td>
|
| 284 |
+
<td>ijk</td>
|
| 285 |
+
<td><b>xyz</b></td>
|
| 286 |
+
</tr>
|
| 287 |
+
</tbody>
|
| 288 |
+
</table>
|