Update README.md
Browse files
README.md
CHANGED
|
@@ -355,16 +355,20 @@ evalplus.evaluate \
|
|
| 355 |
#### Open LLM Leaderboard evaluation scores
|
| 356 |
<table>
|
| 357 |
<tr>
|
| 358 |
-
<
|
| 359 |
-
</
|
| 360 |
-
<
|
| 361 |
-
</
|
| 362 |
-
<
|
| 363 |
-
</
|
| 364 |
-
<
|
| 365 |
-
</
|
|
|
|
|
|
|
| 366 |
</tr>
|
| 367 |
<tr>
|
|
|
|
|
|
|
| 368 |
<td>MMLU (5-shot)
|
| 369 |
</td>
|
| 370 |
<td>80.67
|
|
@@ -434,5 +438,79 @@ evalplus.evaluate \
|
|
| 434 |
<td><strong>99.5%</strong>
|
| 435 |
</td>
|
| 436 |
</tr>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 437 |
</table>
|
| 438 |
|
|
|
|
| 355 |
#### Open LLM Leaderboard evaluation scores
|
| 356 |
<table>
|
| 357 |
<tr>
|
| 358 |
+
<th>Category
|
| 359 |
+
</th>
|
| 360 |
+
<th>Benchmark
|
| 361 |
+
</th>
|
| 362 |
+
<th>Mistral-Small-3.1-24B-Instruct-2503
|
| 363 |
+
</th>
|
| 364 |
+
<th>Mistral-Small-3.1-24B-Instruct-2503-quantized.w4a16<br>(this model)
|
| 365 |
+
</th>
|
| 366 |
+
<th>Recovery
|
| 367 |
+
</th>
|
| 368 |
</tr>
|
| 369 |
<tr>
|
| 370 |
+
<td rowspan="7" ><strong>OpenLLM v1</strong>
|
| 371 |
+
</td>
|
| 372 |
<td>MMLU (5-shot)
|
| 373 |
</td>
|
| 374 |
<td>80.67
|
|
|
|
| 438 |
<td><strong>99.5%</strong>
|
| 439 |
</td>
|
| 440 |
</tr>
|
| 441 |
+
<tr>
|
| 442 |
+
<td rowspan="3" ><strong></strong>
|
| 443 |
+
</td>
|
| 444 |
+
<td>MMLU-Pro (5-shot)
|
| 445 |
+
</td>
|
| 446 |
+
<td>67.25
|
| 447 |
+
</td>
|
| 448 |
+
<td>66.56
|
| 449 |
+
</td>
|
| 450 |
+
<td>99.0%
|
| 451 |
+
</td>
|
| 452 |
+
</tr>
|
| 453 |
+
<tr>
|
| 454 |
+
<td>GPQA CoT main (5-shot)
|
| 455 |
+
</td>
|
| 456 |
+
<td>42.63
|
| 457 |
+
</td>
|
| 458 |
+
<td>47.10
|
| 459 |
+
</td>
|
| 460 |
+
<td>110.5%
|
| 461 |
+
</td>
|
| 462 |
+
</tr>
|
| 463 |
+
<tr>
|
| 464 |
+
<td>GPQA CoT diamond (5-shot)
|
| 465 |
+
</td>
|
| 466 |
+
<td>45.96
|
| 467 |
+
</td>
|
| 468 |
+
<td>44.95
|
| 469 |
+
</td>
|
| 470 |
+
<td>97.80%
|
| 471 |
+
</td>
|
| 472 |
+
</tr>
|
| 473 |
+
<tr>
|
| 474 |
+
<td rowspan="4" ><strong>Coding</strong>
|
| 475 |
+
</td>
|
| 476 |
+
<td>HumanEval pass@1
|
| 477 |
+
</td>
|
| 478 |
+
<td>84.70
|
| 479 |
+
</td>
|
| 480 |
+
<td>84.60
|
| 481 |
+
</td>
|
| 482 |
+
<td>99.9%
|
| 483 |
+
</td>
|
| 484 |
+
</tr>
|
| 485 |
+
<tr>
|
| 486 |
+
<td>HumanEval+ pass@1
|
| 487 |
+
</td>
|
| 488 |
+
<td>79.50
|
| 489 |
+
</td>
|
| 490 |
+
<td>79.90
|
| 491 |
+
</td>
|
| 492 |
+
<td>100.5%
|
| 493 |
+
</td>
|
| 494 |
+
</tr>
|
| 495 |
+
<tr>
|
| 496 |
+
<td>MBPP pass@1
|
| 497 |
+
</td>
|
| 498 |
+
<td>71.10
|
| 499 |
+
</td>
|
| 500 |
+
<td>70.10
|
| 501 |
+
</td>
|
| 502 |
+
<td>98.6%
|
| 503 |
+
</td>
|
| 504 |
+
</tr>
|
| 505 |
+
<tr>
|
| 506 |
+
<td>MBPP+ pass@1
|
| 507 |
+
</td>
|
| 508 |
+
<td>60.60
|
| 509 |
+
</td>
|
| 510 |
+
<td>60.70
|
| 511 |
+
</td>
|
| 512 |
+
<td>100.2%
|
| 513 |
+
</td>
|
| 514 |
+
</tr>
|
| 515 |
</table>
|
| 516 |
|