Hugging Face's logo

Spaces:
steampunque
/
benchlm
Running

App Files Community
steampunque commited on
Commit
b3ee3e9
·
1 Parent(s): 9bd01e7

Update AGIEVALC mathcloze results using answer self grading with LLM

Browse files
Files changed (1) hide show
  1. README.md +31 -26
README.md CHANGED
@@ -15,6 +15,8 @@ Tests are run using a modified llama.cpp server (supporting logprob completion m
15
 
16
  METHODOLOGY:
17
  All CoT and code tests are zero shot.
 
 
18
  All MC tests do two queries, 1 with answers in test order and 2nd with answers circularly shifted 1.
19
  To score a correct answer in MC both queries must answer correctly.
20
  Winogrande using logprob completion (evaluates the probability of a common completion for the two possible cases).
@@ -51,6 +53,8 @@ TESTS:
51
  AGIEVALC - mathcloze, mathqa (Chinese)
52
  MUSR - Multimodal Semantic Reasoning
53
  APPLE - 100 custom Apple Questions
 
 
54
  CODE:
55
  HUMANEVAL - Python
56
  HUMANEVALP - Python, extended test
@@ -93,20 +97,20 @@ GENERAL MODELS:
93
  HUMANEVALX_js | 0.573 | 0.731 | 0.115 | 0.676 | 0.652 | 0.353 | 0.548 | 0.567 | 0.579 | 0.743 | 0.628 | 0.628 | 0.426 | 0.560 | 0.213 | 0.451 | 0.573 | 0.518 | 0.079 | 0.548 | 0.542 | 0.548 | 0.560 | 0.243 | 0.067 | 0.750 | 0.701 | 0.402 | 0.615 | 0.170 | 0.018 | 0.567 | 0.378 | 0.524 | 0.560 | 0.451 | 0.219 | 0.786 | 0.731 | 0.560 | 0.646 | 0.689 | 0.719 | 0.750 | 0.798 | 0.847 | 0.493 | 0.048 | 0.359 |
94
  HUMANEVALX | 0.510 | 0.672 | 0.071 | 0.607 | 0.658 | 0.298 | 0.558 | 0.569 | 0.577 | 0.697 | 0.424 | 0.563 | 0.337 | 0.424 | 0.219 | 0.369 | 0.500 | 0.306 | 0.158 | 0.280 | 0.453 | 0.363 | 0.502 | 0.134 | 0.276 | 0.565 | 0.668 | 0.294 | 0.369 | 0.229 | 0.050 | 0.441 | 0.217 | 0.264 | 0.306 | 0.243 | 0.073 | 0.699 | 0.634 | 0.436 | 0.569 | 0.514 | 0.630 | 0.680 | 0.634 | 0.804 | 0.323 | 0.211 | 0.193 |
95
  CRUXEVAL_input | 0.333 | 0.377 | 0.210 | 0.411 | 0.448 | 0.321 | 0.455 | 0.443 | 0.462 | 0.485 | 0.416 | 0.406 | 0.288 | 0.317 | 0.077 | 0.298 | 0.448 | 0.367 | 0.340 | 0.405 | 0.408 | 0.440 | 0.435 | 0.162 | 0.353 | 0.367 | 0.428 | 0.276 | 0.442 | 0.323 | 0.383 | 0.372 | 0.375 | 0.390 | 0.398 | 0.388 | 0.456 | 0.447 | 0.351 | 0.298 | 0.350 | 0.331 | 0.387 | 0.412 | 0.541 | 0.517 | 0.200 | 0.131 | 0.438 |
96
- CRUXEVAL_output | 0.262 | 0.348 | 0.152 | 0.355 | 0.410 | 0.280 | 0.373 | 0.372 | 0.375 | 0.482 | 0.356 | 0.338 | 0.282 | 0.351 | 0.171 | 0.253 | 0.336 | 0.276 | 0.318 | 0.352 | 0.341 | 0.356 | 0.360 | 0.201 | 0.291 | 0.403 | 0.377 | 0.303 | 0.365 | 0.318 | 0.323 | 0.340 | 0.321 | 0.340 | 0.342 | 0.296 | 0.423 | 0.463 | 0.050 | 0.175 | 0.275 | 0.311 | 0.382 | 0.386 | 0.471 | 0.455 | - | 0.222 | 0.388 |
97
- CRUXEVAL | 0.298 | 0.363 | 0.181 | 0.383 | 0.429 | 0.300 | 0.414 | 0.408 | 0.418 | 0.483 | 0.386 | 0.372 | 0.285 | 0.334 | 0.124 | 0.276 | 0.392 | 0.321 | 0.329 | 0.378 | 0.375 | 0.398 | 0.397 | 0.181 | 0.322 | 0.385 | 0.403 | 0.290 | 0.403 | 0.321 | 0.353 | 0.356 | 0.348 | 0.365 | 0.370 | 0.342 | 0.440 | 0.455 | 0.200 | 0.236 | 0.312 | 0.321 | 0.385 | 0.399 | 0.506 | 0.486 | 0.200 | 0.176 | 0.413 |
98
  CRUXEVALFIM_input | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | 0.195 | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - |
99
  CRUXEVALFIM_output | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | 0.026 | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - |
100
  CRUXEVALFIM | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | 0.110 | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - |
101
- TQA_mc | 0.537 | 0.674 | 0.146 | 0.523 | 0.510 | 0.356 | 0.679 | 0.696 | 0.701 | 0.767 | 0.636 | 0.640 | 0.402 | 0.559 | 0.072 | 0.376 | 0.576 | 0.648 | 0.507 | 0.629 | 0.626 | 0.630 | 0.564 | 0.274 | 0.555 | 0.583 | 0.561 | 0.549 | 0.542 | 0.624 | 0.547 | 0.653 | 0.631 | 0.643 | 0.621 | 0.581 | 0.742 | 0.725 | 0.503 | 0.532 | 0.516 | 0.548 | 0.654 | 0.657 | 0.747 | 0.804 | - | 0.563 | 0.709 |
102
- TQA_tf | 0.510 | 0.647 | 0.381 | 0.410 | 0.431 | 0.510 | 0.675 | 0.719 | 0.692 | 0.725 | 0.484 | 0.457 | 0.421 | 0.473 | 0.250 | 0.510 | 0.343 | 0.593 | 0.504 | 0.643 | 0.552 | 0.563 | 0.512 | 0.063 | 0.566 | 0.569 | 0.572 | 0.548 | 0.479 | 0.548 | 0.536 | 0.578 | 0.532 | 0.541 | 0.483 | 0.487 | 0.670 | 0.686 | 0.435 | 0.602 | 0.414 | 0.300 | 0.574 | 0.568 | 0.706 | 0.731 | - | 0.487 | 0.485 |
103
- TQA | 0.514 | 0.650 | 0.354 | 0.423 | 0.440 | 0.492 | 0.675 | 0.716 | 0.693 | 0.730 | 0.502 | 0.478 | 0.419 | 0.483 | 0.229 | 0.495 | 0.370 | 0.599 | 0.505 | 0.641 | 0.560 | 0.571 | 0.518 | 0.088 | 0.565 | 0.571 | 0.571 | 0.548 | 0.486 | 0.556 | 0.537 | 0.587 | 0.544 | 0.553 | 0.499 | 0.498 | 0.679 | 0.691 | 0.442 | 0.594 | 0.426 | 0.329 | 0.583 | 0.578 | 0.711 | 0.740 | - | 0.496 | 0.511 |
104
- ARC_challenge | 0.709 | 0.825 | 0.374 | 0.809 | 0.819 | 0.671 | 0.874 | 0.881 | 0.882 | 0.897 | 0.835 | 0.853 | 0.614 | 0.742 | 0.063 | 0.599 | 0.744 | 0.812 | 0.732 | 0.794 | 0.769 | 0.766 | 0.776 | 0.342 | 0.706 | 0.825 | 0.775 | 0.688 | 0.773 | 0.797 | 0.819 | 0.838 | 0.808 | 0.833 | 0.813 | 0.802 | 0.888 | 0.911 | 0.818 | 0.796 | 0.750 | 0.777 | 0.843 | 0.851 | 0.911 | 0.934 | - | 0.755 | 0.871 |
105
- ARC_easy | 0.865 | 0.937 | 0.598 | 0.925 | 0.933 | 0.846 | 0.949 | 0.950 | 0.952 | 0.963 | 0.933 | 0.940 | 0.808 | 0.893 | 0.105 | 0.783 | 0.887 | 0.936 | 0.883 | 0.914 | 0.898 | 0.904 | 0.906 | 0.576 | 0.843 | 0.937 | 0.910 | 0.843 | 0.908 | 0.910 | 0.914 | 0.943 | 0.926 | 0.935 | 0.934 | 0.932 | 0.965 | 0.970 | 0.923 | 0.918 | 0.895 | 0.904 | 0.945 | 0.946 | 0.969 | 0.978 | - | 0.899 | 0.962 |
106
- ARC | 0.813 | 0.900 | 0.524 | 0.886 | 0.895 | 0.788 | 0.925 | 0.927 | 0.929 | 0.941 | 0.901 | 0.911 | 0.744 | 0.843 | 0.091 | 0.722 | 0.839 | 0.895 | 0.833 | 0.874 | 0.855 | 0.858 | 0.863 | 0.498 | 0.798 | 0.900 | 0.866 | 0.792 | 0.864 | 0.873 | 0.883 | 0.908 | 0.887 | 0.901 | 0.894 | 0.889 | 0.939 | 0.950 | 0.888 | 0.878 | 0.847 | 0.862 | 0.911 | 0.915 | 0.950 | 0.963 | - | 0.851 | 0.932 |
107
- RACE_high | 0.626 | 0.753 | 0.431 | 0.698 | 0.730 | 0.580 | 0.817 | 0.817 | 0.802 | 0.833 | 0.788 | 0.787 | 0.553 | 0.642 | 0.066 | 0.469 | 0.631 | 0.826 | 0.641 | 0.756 | 0.679 | 0.676 | 0.679 | 0.377 | 0.589 | 0.771 | 0.736 | 0.607 | 0.726 | 0.771 | 0.773 | 0.696 | 0.625 | 0.648 | 0.613 | 0.625 | 0.779 | 0.819 | 0.779 | 0.762 | 0.698 | 0.712 | 0.779 | 0.788 | 0.852 | 0.882 | - | 0.741 | 0.764 |
108
  RACE_middle | 0.704 | 0.806 | 0.463 | 0.777 | 0.793 | 0.610 | 0.860 | 0.860 | 0.849 | 0.883 | 0.816 | 0.825 | 0.631 | 0.735 | 0.089 | 0.579 | 0.706 | 0.863 | 0.708 | 0.817 | 0.747 | 0.744 | 0.734 | 0.396 | 0.680 | 0.825 | 0.800 | 0.696 | 0.782 | 0.807 | 0.834 | 0.750 | 0.697 | 0.722 | 0.706 | 0.692 | 0.832 | 0.861 | 0.827 | 0.811 | 0.775 | 0.776 | 0.841 | 0.853 | 0.887 | 0.923 | - | 0.809 | 0.824 |
109
- RACE | 0.648 | 0.769 | 0.440 | 0.721 | 0.748 | 0.589 | 0.830 | 0.829 | 0.816 | 0.847 | 0.796 | 0.798 | 0.576 | 0.669 | 0.072 | 0.501 | 0.653 | 0.837 | 0.660 | 0.774 | 0.699 | 0.696 | 0.695 | 0.382 | 0.615 | 0.786 | 0.755 | 0.633 | 0.743 | 0.781 | 0.791 | 0.712 | 0.646 | 0.670 | 0.640 | 0.645 | 0.795 | 0.831 | 0.793 | 0.776 | 0.720 | 0.730 | 0.797 | 0.807 | 0.862 | 0.894 | - | 0.761 | 0.781 |
110
  MMLU
111
  abstract_algebra | 0.220 | 0.350 | 0.180 | 0.410 | 0.450 | 0.140 | 0.320 | 0.320 | 0.330 | 0.310 | 0.220 | 0.210 | 0.170 | 0.200 | 0.020 | 0.140 | 0.270 | 0.230 | 0.140 | 0.390 | 0.210 | 0.140 | 0.200 | 0.140 | 0.270 | 0.370 | 0.210 | 0.190 | 0.330 | 0.220 | 0.170 | 0.340 | 0.330 | 0.250 | 0.300 | 0.210 | 0.390 | 0.410 | 0.480 | 0.370 | 0.240 | 0.250 | 0.440 | 0.430 | 0.570 | 0.600 | - | 0.140 | 0.340 |
112
  anatomy | 0.370 | 0.540 | 0.318 | 0.577 | 0.592 | 0.414 | 0.604 | 0.611 | 0.626 | 0.607 | 0.503 | 0.511 | 0.362 | 0.474 | 0.066 | 0.407 | 0.511 | 0.614 | 0.544 | 0.688 | 0.540 | 0.511 | 0.555 | 0.348 | 0.540 | 0.585 | 0.540 | 0.447 | 0.555 | 0.552 | 0.537 | 0.562 | 0.540 | 0.577 | 0.570 | 0.585 | 0.666 | 0.703 | 0.488 | 0.488 | 0.525 | 0.562 | 0.622 | 0.622 | 0.644 | 0.733 | - | 0.477 | 0.607 |
@@ -184,13 +188,13 @@ GENERAL MODELS:
184
  AGIEVALC_jecqaca | - | - | - | - | - | - | - | - | - | - | 0.414 | 0.440 | 0.196 | 0.232 | 0.022 | 0.206 | 0.221 | 0.514 | - | - | - | - | - | - | - | 0.425 | 0.258 | - | 0.247 | 0.223 | 0.273 | - | - | - | - | - | - | - | 0.566 | - | 0.425 | 0.424 | 0.482 | 0.487 | 0.564 | 0.622 | - | 0.185 | 0.266 |
185
  AGIEVALC_jecqakd | - | - | - | - | - | - | - | - | - | - | 0.549 | 0.559 | 0.179 | 0.212 | 0.033 | 0.148 | 0.215 | 0.620 | - | - | - | - | - | - | - | 0.540 | 0.290 | - | 0.304 | 0.281 | 0.275 | - | - | - | - | - | - | - | 0.636 | - | 0.498 | 0.526 | 0.592 | 0.605 | 0.732 | 0.747 | - | 0.242 | 0.288 |
186
  AGIEVALC_logiqa | - | - | - | - | - | - | - | - | - | - | 0.479 | 0.490 | 0.218 | 0.296 | 0.035 | 0.195 | 0.279 | 0.519 | - | - | - | - | - | - | - | 0.442 | 0.313 | - | 0.310 | 0.317 | 0.330 | - | - | - | - | - | - | - | 0.462 | - | 0.399 | 0.405 | 0.497 | 0.500 | 0.565 | 0.588 | - | 0.274 | 0.357 |
187
- AGIEVALC_mathcloze | - | - | - | - | - | - | - | - | - | - | 0.203 | 0.237 | 0.127 | 0.076 | 0.067 | 0.152 | 0.152 | 0.194 | - | - | - | - | - | - | - | 0.262 | 0.177 | - | 0.110 | 0.059 | 0.076 | - | - | - | - | - | - | - | 0.211 | - | 0.279 | 0.254 | 0.279 | 0.279 | 0.211 | 0.237 | - | 0.067 | 0.101 |
188
- AGIEVALC_mathqa | - | - | - | - | - | - | - | - | - | - | 0.621 | 0.648 | 0.343 | 0.404 | 0.281 | 0.401 | 0.465 | 0.662 | - | - | - | - | - | - | - | 0.656 | 0.578 | - | 0.485 | 0.296 | 0.334 | - | - | - | - | - | - | - | 0.543 | - | 0.595 | 0.683 | 0.779 | 0.755 | 0.808 | 0.834 | - | 0.261 | 0.357 |
189
- AGIEVALC_physics | - | - | - | - | - | - | - | - | - | - | 0.396 | 0.425 | 0.166 | 0.229 | 0.034 | 0.166 | 0.189 | 0.436 | - | - | - | - | - | - | - | 0.402 | 0.206 | - | 0.258 | 0.183 | 0.235 | - | - | - | - | - | - | - | 0.494 | - | 0.390 | 0.413 | 0.431 | 0.500 | 0.683 | 0.770 | - | 0.178 | 0.310 |
190
- AGIEVALC | - | - | - | - | - | - | - | - | - | - | 0.580 | 0.597 | 0.255 | 0.317 | 0.056 | 0.243 | 0.314 | 0.632 | - | - | - | - | - | - | - | 0.563 | 0.399 | - | 0.399 | 0.322 | 0.366 | - | - | - | - | - | - | - | 0.605 | - | 0.522 | 0.542 | 0.613 | 0.623 | 0.699 | 0.716 | - | 0.297 | 0.387 |
191
  BBH
192
  boolean_expressions | 0.720 | 0.740 | 0.544 | 0.860 | 0.876 | 0.556 | 0.764 | 0.776 | 0.768 | 0.460 | 0.848 | 0.868 | 0.812 | 0.856 | 0.724 | 0.752 | 0.700 | 0.688 | 0.796 | 0.804 | 0.824 | 0.832 | 0.844 | 0.460 | 0.480 | 0.896 | 0.728 | 0.764 | 0.780 | 0.824 | 0.664 | 0.816 | 0.848 | 0.800 | 0.852 | 0.832 | 0.696 | 0.936 | 0.808 | 0.776 | 0.756 | 0.796 | 0.864 | 0.880 | 0.888 | 0.808 | - | 0.720 | 0.540 |
193
- causal_judgement | 0.588 | 0.598 | 0.550 | 0.577 | 0.582 | 0.524 | 0.609 | 0.604 | 0.598 | 0.604 | 0.550 | 0.550 | 0.566 | 0.598 | 0.566 | 0.604 | 0.588 | 0.689 | 0.540 | 0.513 | 0.545 | 0.518 | 0.540 | 0.502 | 0.518 | 0.577 | 0.593 | 0.588 | 0.625 | 0.614 | 0.604 | 0.556 | 0.508 | 0.598 | 0.588 | 0.593 | 0.588 | 0.647 | 0.625 | 0.582 | 0.497 | 0.529 | 0.522 | 0.513 | 0.647 | 0.700 | - | 0.636 | 0.641 |
194
  date_understanding | 0.576 | 0.752 | 0.324 | 0.668 | 0.748 | 0.592 | 0.780 | 0.764 | 0.748 | 0.788 | 0.580 | 0.572 | 0.560 | 0.684 | 0.284 | 0.628 | 0.660 | 0.832 | 0.700 | 0.660 | 0.732 | 0.724 | 0.716 | 0.400 | 0.664 | 0.728 | 0.772 | 0.548 | 0.668 | 0.592 | 0.608 | 0.644 | 0.464 | 0.568 | 0.696 | 0.576 | 0.780 | 0.932 | 0.544 | 0.544 | 0.616 | 0.648 | 0.764 | 0.740 | 0.856 | 0.872 | - | 0.556 | 0.724 |
195
  disambiguation_qa | 0.612 | 0.588 | 0.400 | 0.712 | 0.668 | 0.532 | 0.688 | 0.652 | 0.660 | 0.720 | 0.584 | 0.636 | 0.628 | 0.640 | 0.380 | 0.616 | 0.648 | 0.732 | 0.584 | 0.536 | 0.552 | 0.540 | 0.516 | 0.424 | 0.472 | 0.688 | 0.644 | 0.600 | 0.596 | 0.728 | 0.704 | 0.604 | 0.640 | 0.592 | 0.720 | 0.752 | 0.692 | 0.768 | 0.660 | 0.696 | 0.544 | 0.556 | 0.656 | 0.636 | 0.764 | 0.780 | - | 0.576 | 0.640 |
196
  dyck_languages | 0.596 | 0.664 | 0.424 | 0.704 | 0.712 | 0.476 | 0.752 | 0.720 | 0.728 | 0.600 | 0.516 | 0.544 | 0.560 | 0.704 | 0.356 | 0.700 | 0.664 | 0.728 | 0.792 | 0.228 | 0.832 | 0.724 | 0.796 | 0.536 | 0.680 | 0.736 | 0.756 | 0.744 | 0.712 | 0.664 | 0.732 | 0.752 | 0.532 | 0.424 | 0.580 | 0.468 | 0.532 | 0.776 | 0.756 | 0.576 | 0.596 | 0.628 | 0.868 | 0.836 | 0.648 | 0.820 | - | 0.684 | 0.572 |
@@ -239,24 +243,25 @@ GENERAL MODELS:
239
  psychology | 0.502 | 0.642 | 0.258 | 0.621 | 0.645 | 0.494 | 0.666 | 0.657 | 0.647 | 0.692 | 0.586 | 0.602 | 0.436 | 0.531 | 0.184 | 0.448 | 0.543 | 0.572 | 0.560 | 0.581 | 0.624 | 0.601 | 0.637 | 0.317 | 0.518 | 0.604 | 0.595 | - | 0.588 | 0.543 | 0.552 | 0.626 | 0.573 | 0.583 | 0.621 | 0.572 | 0.676 | 0.759 | 0.595 | 0.536 | 0.526 | 0.563 | 0.636 | 0.644 | 0.721 | 0.749 | - | 0.525 | - |
240
  MMLUPRO | 0.416 | 0.554 | 0.186 | 0.552 | 0.517 | 0.326 | 0.524 | 0.553 | 0.528 | 0.568 | 0.471 | 0.480 | 0.298 | 0.395 | 0.145 | 0.324 | 0.410 | 0.432 | 0.404 | 0.475 | 0.494 | 0.491 | 0.499 | 0.215 | 0.419 | 0.539 | 0.458 | 0.453 | 0.436 | 0.376 | 0.382 | 0.475 | 0.405 | 0.451 | 0.482 | 0.408 | 0.470 | 0.719 | 0.475 | 0.368 | 0.430 | 0.457 | 0.564 | 0.575 | 0.649 | 0.671 | - | 0.326 | 0.509 |
241
  CATEGORIES
242
- REASONING | 0.658 | 0.774 | 0.367 | 0.713 | 0.738 | 0.570 | 0.782 | 0.800 | 0.788 | 0.814 | 0.804 | 0.811 | 0.592 | 0.702 | 0.096 | 0.561 | 0.703 | 0.845 | 0.684 | 0.754 | 0.702 | 0.708 | 0.713 | 0.352 | 0.606 | 0.791 | 0.779 | 0.628 | 0.730 | 0.785 | 0.799 | 0.744 | 0.713 | 0.741 | 0.724 | 0.691 | 0.805 | 0.809 | 0.755 | 0.747 | 0.689 | 0.719 | 0.805 | 0.809 | 0.850 | 0.874 | 0.867 | 0.784 | 0.806 |
243
- UNDERSTANDING | 0.548 | 0.652 | 0.366 | 0.644 | 0.670 | 0.538 | 0.708 | 0.712 | 0.707 | 0.742 | 0.661 | 0.670 | 0.511 | 0.598 | 0.116 | 0.481 | 0.599 | 0.685 | 0.602 | 0.680 | 0.622 | 0.618 | 0.631 | 0.330 | 0.579 | 0.672 | 0.633 | 0.563 | 0.633 | 0.644 | 0.651 | 0.661 | 0.629 | 0.629 | 0.614 | 0.622 | 0.727 | 0.728 | 0.674 | 0.649 | 0.605 | 0.613 | 0.692 | 0.696 | 0.761 | 0.793 | 0.779 | 0.617 | 0.713 |
244
  LANGUAGE | 0.613 | 0.680 | 0.524 | 0.688 | 0.692 | 0.624 | 0.733 | 0.732 | 0.735 | 0.755 | 0.786 | 0.783 | 0.746 | 0.799 | 0.665 | 0.732 | 0.790 | 0.732 | 0.710 | 0.729 | 0.740 | 0.738 | 0.747 | 0.610 | 0.705 | 0.715 | 0.776 | 0.766 | 0.714 | 0.744 | 0.733 | 0.613 | 0.638 | 0.618 | 0.677 | 0.613 | 0.632 | 0.750 | 0.735 | 0.564 | 0.685 | 0.682 | 0.722 | 0.724 | 0.769 | 0.781 | 0.780 | 0.654 | 0.708 |
245
- KNOWLEDGE | 0.516 | 0.627 | 0.354 | 0.476 | 0.496 | 0.505 | 0.677 | 0.710 | 0.689 | 0.733 | 0.542 | 0.543 | 0.406 | 0.470 | 0.266 | 0.358 | 0.404 | 0.588 | 0.536 | 0.633 | 0.568 | 0.571 | 0.547 | 0.066 | 0.536 | 0.580 | 0.526 | 0.582 | 0.511 | 0.607 | 0.581 | 0.581 | 0.546 | 0.546 | 0.517 | 0.519 | 0.663 | 0.678 | 0.530 | 0.585 | 0.469 | 0.426 | 0.595 | 0.597 | 0.693 | 0.725 | 0.629 | 0.545 | 0.521 |
246
  COT | 0.438 | 0.561 | 0.220 | 0.552 | 0.530 | 0.350 | 0.548 | 0.568 | 0.550 | 0.582 | 0.485 | 0.500 | 0.336 | 0.431 | 0.188 | 0.365 | 0.445 | 0.505 | 0.446 | 0.488 | 0.522 | 0.521 | 0.530 | 0.255 | 0.446 | 0.545 | 0.479 | 0.498 | 0.466 | 0.416 | 0.424 | 0.502 | 0.437 | 0.474 | 0.506 | 0.440 | 0.503 | 0.725 | 0.492 | 0.398 | 0.443 | 0.462 | 0.570 | 0.581 | 0.653 | 0.684 | - | 0.377 | 0.563 |
247
- MATHCOT | 0.731 | 0.819 | 0.369 | 0.745 | 0.752 | 0.480 | 0.728 | 0.733 | 0.735 | 0.740 | 0.676 | 0.672 | 0.544 | 0.635 | 0.346 | 0.568 | 0.644 | 0.705 | 0.622 | 0.680 | 0.708 | 0.721 | 0.728 | 0.386 | 0.647 | 0.800 | 0.743 | 0.493 | 0.649 | 0.544 | 0.575 | 0.695 | 0.612 | 0.591 | 0.767 | 0.638 | 0.665 | 0.919 | 0.666 | 0.548 | 0.663 | 0.691 | 0.814 | 0.821 | 0.858 | 0.891 | 0.956 | 0.534 | 0.651 |
248
- CODE | 0.416 | 0.498 | 0.176 | 0.499 | 0.534 | 0.331 | 0.487 | 0.483 | 0.495 | 0.568 | 0.456 | 0.475 | 0.339 | 0.410 | 0.210 | 0.356 | 0.466 | 0.344 | 0.269 | 0.411 | 0.440 | 0.439 | 0.463 | 0.217 | 0.366 | 0.498 | 0.514 | 0.321 | 0.326 | 0.324 | 0.316 | 0.430 | 0.372 | 0.389 | 0.427 | 0.376 | 0.350 | 0.568 | 0.390 | 0.346 | 0.437 | 0.445 | 0.510 | 0.528 | 0.578 | 0.612 | 0.303 | 0.233 | 0.368 |
249
  DISCIPLINES
250
- NLP | 0.620 | 0.720 | 0.408 | 0.647 | 0.670 | 0.568 | 0.751 | 0.767 | 0.755 | 0.786 | 0.729 | 0.728 | 0.588 | 0.667 | 0.263 | 0.549 | 0.648 | 0.774 | 0.650 | 0.713 | 0.675 | 0.678 | 0.677 | 0.329 | 0.609 | 0.722 | 0.723 | 0.642 | 0.685 | 0.739 | 0.737 | 0.681 | 0.655 | 0.667 | 0.647 | 0.637 | 0.744 | 0.755 | 0.693 | 0.687 | 0.632 | 0.630 | 0.731 | 0.734 | 0.791 | 0.818 | 0.814 | 0.712 | 0.725 |
251
- MATH | 0.597 | 0.712 | 0.294 | 0.674 | 0.659 | 0.398 | 0.634 | 0.650 | 0.637 | 0.653 | 0.587 | 0.594 | 0.444 | 0.542 | 0.258 | 0.480 | 0.561 | 0.638 | 0.528 | 0.594 | 0.612 | 0.613 | 0.629 | 0.318 | 0.556 | 0.707 | 0.630 | 0.451 | 0.556 | 0.482 | 0.506 | 0.611 | 0.527 | 0.525 | 0.646 | 0.543 | 0.585 | 0.817 | 0.609 | 0.493 | 0.573 | 0.596 | 0.736 | 0.742 | 0.793 | 0.836 | 0.956 | 0.455 | 0.608 |
252
- SCIENCE | 0.606 | 0.726 | 0.350 | 0.741 | 0.713 | 0.555 | 0.735 | 0.749 | 0.739 | 0.769 | 0.690 | 0.698 | 0.481 | 0.576 | 0.093 | 0.481 | 0.579 | 0.663 | 0.608 | 0.685 | 0.668 | 0.668 | 0.676 | 0.346 | 0.605 | 0.716 | 0.621 | 0.673 | 0.618 | 0.611 | 0.619 | 0.701 | 0.658 | 0.685 | 0.696 | 0.660 | 0.697 | 0.845 | 0.674 | 0.618 | 0.629 | 0.657 | 0.738 | 0.748 | 0.815 | 0.806 | - | 0.571 | 0.731 |
253
  ENGINEERING | 0.349 | 0.429 | 0.166 | 0.464 | 0.453 | 0.280 | 0.412 | 0.426 | 0.426 | 0.438 | 0.334 | 0.333 | 0.224 | 0.303 | 0.108 | 0.253 | 0.289 | 0.333 | 0.332 | 0.386 | 0.412 | 0.404 | 0.397 | 0.169 | 0.323 | 0.407 | 0.346 | 0.393 | 0.339 | 0.267 | 0.272 | 0.305 | 0.300 | 0.319 | 0.323 | 0.308 | 0.371 | 0.595 | 0.388 | 0.283 | 0.315 | 0.325 | 0.443 | 0.444 | 0.530 | 0.590 | - | 0.199 | 0.586 |
254
  MEDICINE | 0.379 | 0.504 | 0.216 | 0.524 | 0.540 | 0.400 | 0.591 | 0.590 | 0.595 | 0.648 | 0.521 | 0.530 | 0.346 | 0.464 | 0.069 | 0.347 | 0.469 | 0.515 | 0.544 | 0.620 | 0.568 | 0.572 | 0.577 | 0.243 | 0.496 | 0.543 | 0.512 | 0.447 | 0.541 | 0.485 | 0.503 | 0.558 | 0.507 | 0.525 | 0.537 | 0.501 | 0.633 | 0.672 | 0.510 | 0.482 | 0.459 | 0.478 | 0.574 | 0.580 | 0.655 | 0.702 | 0.598 | 0.457 | 0.635 |
255
- HUMANITIES | 0.472 | 0.594 | 0.291 | 0.572 | 0.615 | 0.485 | 0.643 | 0.652 | 0.645 | 0.679 | 0.586 | 0.610 | 0.416 | 0.521 | 0.094 | 0.395 | 0.517 | 0.590 | 0.545 | 0.624 | 0.571 | 0.563 | 0.578 | 0.292 | 0.529 | 0.609 | 0.552 | 0.536 | 0.552 | 0.551 | 0.559 | 0.605 | 0.567 | 0.567 | 0.588 | 0.567 | 0.671 | 0.740 | 0.597 | 0.544 | 0.527 | 0.533 | 0.629 | 0.638 | 0.716 | 0.742 | 0.600 | 0.524 | 0.660 |
256
  BUSINESS | 0.536 | 0.657 | 0.252 | 0.679 | 0.655 | 0.450 | 0.660 | 0.697 | 0.678 | 0.709 | 0.623 | 0.637 | 0.408 | 0.523 | 0.115 | 0.428 | 0.530 | 0.594 | 0.537 | 0.626 | 0.592 | 0.582 | 0.598 | 0.251 | 0.517 | 0.667 | 0.564 | 0.466 | 0.565 | 0.514 | 0.526 | 0.632 | 0.575 | 0.589 | 0.637 | 0.604 | 0.636 | 0.801 | 0.644 | 0.566 | 0.565 | 0.596 | 0.701 | 0.710 | 0.759 | 0.802 | - | 0.479 | 0.652 |
257
- LAW | 0.316 | 0.419 | 0.200 | 0.396 | 0.427 | 0.302 | 0.489 | 0.485 | 0.483 | 0.524 | 0.402 | 0.429 | 0.280 | 0.344 | 0.075 | 0.258 | 0.348 | 0.385 | 0.374 | 0.412 | 0.380 | 0.397 | 0.406 | 0.175 | 0.344 | 0.420 | 0.365 | 0.370 | 0.367 | 0.395 | 0.399 | 0.427 | 0.393 | 0.399 | 0.392 | 0.310 | 0.498 | 0.541 | 0.446 | 0.376 | 0.374 | 0.383 | 0.451 | 0.456 | 0.541 | 0.604 | - | 0.355 | 0.465 |
258
  COMPOSITE AVERAGE
259
- AVG | 0.561 | 0.668 | 0.342 | 0.627 | 0.641 | 0.502 | 0.688 | 0.701 | 0.692 | 0.724 | 0.651 | 0.653 | 0.493 | 0.581 | 0.197 | 0.476 | 0.576 | 0.671 | 0.582 | 0.651 | 0.622 | 0.623 | 0.629 | 0.306 | 0.561 | 0.668 | 0.634 | 0.578 | 0.598 | 0.610 | 0.616 | 0.634 | 0.595 | 0.606 | 0.616 | 0.585 | 0.674 | 0.748 | 0.632 | 0.597 | 0.578 | 0.586 | 0.685 | 0.691 | 0.753 | 0.782 | 0.779 | 0.578 | 0.674 |
 
260
 
261
  CODE MODELS:
262
 
 
15
 
16
  METHODOLOGY:
17
  All CoT and code tests are zero shot.
18
+ Math CoT test such as GSM8K, APPLE, MATH etc. are self graded against correct answer using LLM under test
19
+ If self grade does not work reliably (such as with very small model) the result is zeroed to mark invalid test.
20
  All MC tests do two queries, 1 with answers in test order and 2nd with answers circularly shifted 1.
21
  To score a correct answer in MC both queries must answer correctly.
22
  Winogrande using logprob completion (evaluates the probability of a common completion for the two possible cases).
 
53
  AGIEVALC - mathcloze, mathqa (Chinese)
54
  MUSR - Multimodal Semantic Reasoning
55
  APPLE - 100 custom Apple Questions
56
+ MATH:
57
+ MATH1..MATH5 - MATH Datasets level 1 through 5 (Hendrycks et al.)
58
  CODE:
59
  HUMANEVAL - Python
60
  HUMANEVALP - Python, extended test
 
97
  HUMANEVALX_js | 0.573 | 0.731 | 0.115 | 0.676 | 0.652 | 0.353 | 0.548 | 0.567 | 0.579 | 0.743 | 0.628 | 0.628 | 0.426 | 0.560 | 0.213 | 0.451 | 0.573 | 0.518 | 0.079 | 0.548 | 0.542 | 0.548 | 0.560 | 0.243 | 0.067 | 0.750 | 0.701 | 0.402 | 0.615 | 0.170 | 0.018 | 0.567 | 0.378 | 0.524 | 0.560 | 0.451 | 0.219 | 0.786 | 0.731 | 0.560 | 0.646 | 0.689 | 0.719 | 0.750 | 0.798 | 0.847 | 0.493 | 0.048 | 0.359 |
98
  HUMANEVALX | 0.510 | 0.672 | 0.071 | 0.607 | 0.658 | 0.298 | 0.558 | 0.569 | 0.577 | 0.697 | 0.424 | 0.563 | 0.337 | 0.424 | 0.219 | 0.369 | 0.500 | 0.306 | 0.158 | 0.280 | 0.453 | 0.363 | 0.502 | 0.134 | 0.276 | 0.565 | 0.668 | 0.294 | 0.369 | 0.229 | 0.050 | 0.441 | 0.217 | 0.264 | 0.306 | 0.243 | 0.073 | 0.699 | 0.634 | 0.436 | 0.569 | 0.514 | 0.630 | 0.680 | 0.634 | 0.804 | 0.323 | 0.211 | 0.193 |
99
  CRUXEVAL_input | 0.333 | 0.377 | 0.210 | 0.411 | 0.448 | 0.321 | 0.455 | 0.443 | 0.462 | 0.485 | 0.416 | 0.406 | 0.288 | 0.317 | 0.077 | 0.298 | 0.448 | 0.367 | 0.340 | 0.405 | 0.408 | 0.440 | 0.435 | 0.162 | 0.353 | 0.367 | 0.428 | 0.276 | 0.442 | 0.323 | 0.383 | 0.372 | 0.375 | 0.390 | 0.398 | 0.388 | 0.456 | 0.447 | 0.351 | 0.298 | 0.350 | 0.331 | 0.387 | 0.412 | 0.541 | 0.517 | 0.200 | 0.131 | 0.438 |
100
+ CRUXEVAL_output | 0.262 | 0.348 | 0.152 | 0.355 | 0.410 | 0.280 | 0.373 | 0.372 | 0.375 | 0.482 | 0.356 | 0.338 | 0.282 | 0.351 | 0.171 | 0.253 | 0.336 | 0.276 | 0.318 | 0.352 | 0.341 | 0.356 | 0.360 | 0.201 | 0.291 | 0.403 | 0.377 | 0.303 | 0.365 | 0.318 | 0.323 | 0.340 | 0.321 | 0.340 | 0.342 | 0.296 | 0.423 | 0.463 | 0.050 | 0.175 | 0.275 | 0.311 | 0.382 | 0.386 | 0.471 | 0.455 | 0.368 | 0.222 | 0.388 |
101
+ CRUXEVAL | 0.298 | 0.363 | 0.181 | 0.383 | 0.429 | 0.300 | 0.414 | 0.408 | 0.418 | 0.483 | 0.386 | 0.372 | 0.285 | 0.334 | 0.124 | 0.276 | 0.392 | 0.321 | 0.329 | 0.378 | 0.375 | 0.398 | 0.397 | 0.181 | 0.322 | 0.385 | 0.403 | 0.290 | 0.403 | 0.321 | 0.353 | 0.356 | 0.348 | 0.365 | 0.370 | 0.342 | 0.440 | 0.455 | 0.200 | 0.236 | 0.312 | 0.321 | 0.385 | 0.399 | 0.506 | 0.486 | 0.284 | 0.176 | 0.413 |
102
  CRUXEVALFIM_input | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | 0.195 | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - |
103
  CRUXEVALFIM_output | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | 0.026 | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - |
104
  CRUXEVALFIM | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | 0.110 | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - |
105
+ TQA_mc | 0.537 | 0.674 | 0.146 | 0.523 | 0.510 | 0.356 | 0.679 | 0.696 | 0.701 | 0.767 | 0.636 | 0.640 | 0.402 | 0.559 | 0.072 | 0.376 | 0.576 | 0.648 | 0.507 | 0.629 | 0.626 | 0.630 | 0.564 | 0.274 | 0.555 | 0.583 | 0.561 | 0.549 | 0.542 | 0.624 | 0.547 | 0.653 | 0.631 | 0.643 | 0.621 | 0.581 | 0.742 | 0.725 | 0.503 | 0.532 | 0.516 | 0.548 | 0.654 | 0.657 | 0.747 | 0.804 | 0.795 | 0.563 | 0.709 |
106
+ TQA_tf | 0.510 | 0.647 | 0.381 | 0.410 | 0.431 | 0.510 | 0.675 | 0.719 | 0.692 | 0.725 | 0.484 | 0.457 | 0.421 | 0.473 | 0.250 | 0.510 | 0.343 | 0.593 | 0.504 | 0.643 | 0.552 | 0.563 | 0.512 | 0.063 | 0.566 | 0.569 | 0.572 | 0.548 | 0.479 | 0.548 | 0.536 | 0.578 | 0.532 | 0.541 | 0.483 | 0.487 | 0.670 | 0.686 | 0.435 | 0.602 | 0.414 | 0.300 | 0.574 | 0.568 | 0.706 | 0.731 | 0.523 | 0.487 | 0.485 |
107
+ TQA | 0.514 | 0.650 | 0.354 | 0.423 | 0.440 | 0.492 | 0.675 | 0.716 | 0.693 | 0.730 | 0.502 | 0.478 | 0.419 | 0.483 | 0.229 | 0.495 | 0.370 | 0.599 | 0.505 | 0.641 | 0.560 | 0.571 | 0.518 | 0.088 | 0.565 | 0.571 | 0.571 | 0.548 | 0.486 | 0.556 | 0.537 | 0.587 | 0.544 | 0.553 | 0.499 | 0.498 | 0.679 | 0.691 | 0.442 | 0.594 | 0.426 | 0.329 | 0.583 | 0.578 | 0.711 | 0.740 | 0.554 | 0.496 | 0.511 |
108
+ ARC_challenge | 0.709 | 0.825 | 0.374 | 0.809 | 0.819 | 0.671 | 0.874 | 0.881 | 0.882 | 0.897 | 0.835 | 0.853 | 0.614 | 0.742 | 0.063 | 0.599 | 0.744 | 0.812 | 0.732 | 0.794 | 0.769 | 0.766 | 0.776 | 0.342 | 0.706 | 0.825 | 0.775 | 0.688 | 0.773 | 0.797 | 0.819 | 0.838 | 0.808 | 0.833 | 0.813 | 0.802 | 0.888 | 0.911 | 0.818 | 0.796 | 0.750 | 0.777 | 0.843 | 0.851 | 0.911 | 0.934 | 0.917 | 0.755 | 0.871 |
109
+ ARC_easy | 0.865 | 0.937 | 0.598 | 0.925 | 0.933 | 0.846 | 0.949 | 0.950 | 0.952 | 0.963 | 0.933 | 0.940 | 0.808 | 0.893 | 0.105 | 0.783 | 0.887 | 0.936 | 0.883 | 0.914 | 0.898 | 0.904 | 0.906 | 0.576 | 0.843 | 0.937 | 0.910 | 0.843 | 0.908 | 0.910 | 0.914 | 0.943 | 0.926 | 0.935 | 0.934 | 0.932 | 0.965 | 0.970 | 0.923 | 0.918 | 0.895 | 0.904 | 0.945 | 0.946 | 0.969 | 0.978 | 0.975 | 0.899 | 0.962 |
110
+ ARC | 0.813 | 0.900 | 0.524 | 0.886 | 0.895 | 0.788 | 0.925 | 0.927 | 0.929 | 0.941 | 0.901 | 0.911 | 0.744 | 0.843 | 0.091 | 0.722 | 0.839 | 0.895 | 0.833 | 0.874 | 0.855 | 0.858 | 0.863 | 0.498 | 0.798 | 0.900 | 0.866 | 0.792 | 0.864 | 0.873 | 0.883 | 0.908 | 0.887 | 0.901 | 0.894 | 0.889 | 0.939 | 0.950 | 0.888 | 0.878 | 0.847 | 0.862 | 0.911 | 0.915 | 0.950 | 0.963 | 0.956 | 0.851 | 0.932 |
111
+ RACE_high | 0.626 | 0.753 | 0.431 | 0.698 | 0.730 | 0.580 | 0.817 | 0.817 | 0.802 | 0.833 | 0.788 | 0.787 | 0.553 | 0.642 | 0.066 | 0.469 | 0.631 | 0.826 | 0.641 | 0.756 | 0.679 | 0.676 | 0.679 | 0.377 | 0.589 | 0.771 | 0.736 | 0.607 | 0.726 | 0.771 | 0.773 | 0.696 | 0.625 | 0.648 | 0.613 | 0.625 | 0.779 | 0.819 | 0.779 | 0.762 | 0.698 | 0.712 | 0.779 | 0.788 | 0.852 | 0.882 | 0.871 | 0.741 | 0.764 |
112
  RACE_middle | 0.704 | 0.806 | 0.463 | 0.777 | 0.793 | 0.610 | 0.860 | 0.860 | 0.849 | 0.883 | 0.816 | 0.825 | 0.631 | 0.735 | 0.089 | 0.579 | 0.706 | 0.863 | 0.708 | 0.817 | 0.747 | 0.744 | 0.734 | 0.396 | 0.680 | 0.825 | 0.800 | 0.696 | 0.782 | 0.807 | 0.834 | 0.750 | 0.697 | 0.722 | 0.706 | 0.692 | 0.832 | 0.861 | 0.827 | 0.811 | 0.775 | 0.776 | 0.841 | 0.853 | 0.887 | 0.923 | - | 0.809 | 0.824 |
113
+ RACE | 0.648 | 0.769 | 0.440 | 0.721 | 0.748 | 0.589 | 0.830 | 0.829 | 0.816 | 0.847 | 0.796 | 0.798 | 0.576 | 0.669 | 0.072 | 0.501 | 0.653 | 0.837 | 0.660 | 0.774 | 0.699 | 0.696 | 0.695 | 0.382 | 0.615 | 0.786 | 0.755 | 0.633 | 0.743 | 0.781 | 0.791 | 0.712 | 0.646 | 0.670 | 0.640 | 0.645 | 0.795 | 0.831 | 0.793 | 0.776 | 0.720 | 0.730 | 0.797 | 0.807 | 0.862 | 0.894 | 0.871 | 0.761 | 0.781 |
114
  MMLU
115
  abstract_algebra | 0.220 | 0.350 | 0.180 | 0.410 | 0.450 | 0.140 | 0.320 | 0.320 | 0.330 | 0.310 | 0.220 | 0.210 | 0.170 | 0.200 | 0.020 | 0.140 | 0.270 | 0.230 | 0.140 | 0.390 | 0.210 | 0.140 | 0.200 | 0.140 | 0.270 | 0.370 | 0.210 | 0.190 | 0.330 | 0.220 | 0.170 | 0.340 | 0.330 | 0.250 | 0.300 | 0.210 | 0.390 | 0.410 | 0.480 | 0.370 | 0.240 | 0.250 | 0.440 | 0.430 | 0.570 | 0.600 | - | 0.140 | 0.340 |
116
  anatomy | 0.370 | 0.540 | 0.318 | 0.577 | 0.592 | 0.414 | 0.604 | 0.611 | 0.626 | 0.607 | 0.503 | 0.511 | 0.362 | 0.474 | 0.066 | 0.407 | 0.511 | 0.614 | 0.544 | 0.688 | 0.540 | 0.511 | 0.555 | 0.348 | 0.540 | 0.585 | 0.540 | 0.447 | 0.555 | 0.552 | 0.537 | 0.562 | 0.540 | 0.577 | 0.570 | 0.585 | 0.666 | 0.703 | 0.488 | 0.488 | 0.525 | 0.562 | 0.622 | 0.622 | 0.644 | 0.733 | - | 0.477 | 0.607 |
 
188
  AGIEVALC_jecqaca | - | - | - | - | - | - | - | - | - | - | 0.414 | 0.440 | 0.196 | 0.232 | 0.022 | 0.206 | 0.221 | 0.514 | - | - | - | - | - | - | - | 0.425 | 0.258 | - | 0.247 | 0.223 | 0.273 | - | - | - | - | - | - | - | 0.566 | - | 0.425 | 0.424 | 0.482 | 0.487 | 0.564 | 0.622 | - | 0.185 | 0.266 |
189
  AGIEVALC_jecqakd | - | - | - | - | - | - | - | - | - | - | 0.549 | 0.559 | 0.179 | 0.212 | 0.033 | 0.148 | 0.215 | 0.620 | - | - | - | - | - | - | - | 0.540 | 0.290 | - | 0.304 | 0.281 | 0.275 | - | - | - | - | - | - | - | 0.636 | - | 0.498 | 0.526 | 0.592 | 0.605 | 0.732 | 0.747 | - | 0.242 | 0.288 |
190
  AGIEVALC_logiqa | - | - | - | - | - | - | - | - | - | - | 0.479 | 0.490 | 0.218 | 0.296 | 0.035 | 0.195 | 0.279 | 0.519 | - | - | - | - | - | - | - | 0.442 | 0.313 | - | 0.310 | 0.317 | 0.330 | - | - | - | - | - | - | - | 0.462 | - | 0.399 | 0.405 | 0.497 | 0.500 | 0.565 | 0.588 | - | 0.274 | 0.357 |
191
+ AGIEVALC_mathcloze | - | - | - | - | - | - | - | - | - | - | 0.491 | 0.542 | 0.186 | 0.237 | - | 0.288 | 0.415 | 0.576 | - | - | - | - | - | - | - | 0.652 | 0.508 | - | 0.288 | 0.152 | 0.245 | - | - | - | - | - | - | - | 0.449 | - | 0.508 | 0.440 | 0.694 | 0.686 | 0.737 | 0.805 | 0.864 | 0.110 | 0.618 |
192
+ AGIEVALC_mathqa | - | - | - | - | - | - | - | - | - | - | 0.621 | 0.648 | 0.343 | 0.404 | 0.281 | 0.401 | 0.465 | 0.662 | - | - | - | - | - | - | - | 0.656 | 0.578 | - | 0.485 | 0.296 | 0.334 | - | - | - | - | - | - | - | 0.543 | - | 0.595 | 0.683 | 0.779 | 0.755 | 0.808 | 0.834 | 0.828 | 0.261 | 0.357 |
193
+ AGIEVALC_physics | - | - | - | - | - | - | - | - | - | - | 0.396 | 0.425 | 0.166 | 0.229 | 0.034 | 0.166 | 0.189 | 0.436 | - | - | - | - | - | - | - | 0.402 | 0.206 | - | 0.258 | 0.183 | 0.235 | - | - | - | - | - | - | - | 0.494 | - | 0.390 | 0.413 | 0.431 | 0.500 | 0.683 | 0.770 | 0.741 | 0.178 | 0.310 |
194
+ AGIEVALC | - | - | - | - | - | - | - | - | - | - | 0.589 | 0.607 | 0.257 | 0.322 | 0.076 | 0.248 | 0.322 | 0.645 | - | - | - | - | - | - | - | 0.576 | 0.409 | - | 0.404 | 0.325 | 0.371 | - | - | - | - | - | - | - | 0.612 | - | 0.529 | 0.548 | 0.627 | 0.636 | 0.716 | 0.734 | 0.811 | 0.298 | 0.403 |
195
  BBH
196
  boolean_expressions | 0.720 | 0.740 | 0.544 | 0.860 | 0.876 | 0.556 | 0.764 | 0.776 | 0.768 | 0.460 | 0.848 | 0.868 | 0.812 | 0.856 | 0.724 | 0.752 | 0.700 | 0.688 | 0.796 | 0.804 | 0.824 | 0.832 | 0.844 | 0.460 | 0.480 | 0.896 | 0.728 | 0.764 | 0.780 | 0.824 | 0.664 | 0.816 | 0.848 | 0.800 | 0.852 | 0.832 | 0.696 | 0.936 | 0.808 | 0.776 | 0.756 | 0.796 | 0.864 | 0.880 | 0.888 | 0.808 | - | 0.720 | 0.540 |
197
+ causal_judgement | 0.588 | 0.598 | 0.550 | 0.577 | 0.582 | 0.524 | 0.609 | 0.604 | 0.598 | 0.604 | 0.550 | 0.550 | 0.566 | 0.598 | 0.566 | 0.604 | 0.588 | 0.689 | 0.540 | 0.513 | 0.545 | 0.518 | 0.540 | 0.502 | 0.518 | 0.577 | 0.593 | 0.588 | 0.625 | 0.614 | 0.604 | 0.556 | 0.508 | 0.598 | 0.588 | 0.593 | 0.588 | 0.647 | 0.625 | 0.582 | 0.497 | 0.529 | 0.508 | 0.513 | 0.647 | 0.700 | - | 0.636 | 0.641 |
198
  date_understanding | 0.576 | 0.752 | 0.324 | 0.668 | 0.748 | 0.592 | 0.780 | 0.764 | 0.748 | 0.788 | 0.580 | 0.572 | 0.560 | 0.684 | 0.284 | 0.628 | 0.660 | 0.832 | 0.700 | 0.660 | 0.732 | 0.724 | 0.716 | 0.400 | 0.664 | 0.728 | 0.772 | 0.548 | 0.668 | 0.592 | 0.608 | 0.644 | 0.464 | 0.568 | 0.696 | 0.576 | 0.780 | 0.932 | 0.544 | 0.544 | 0.616 | 0.648 | 0.764 | 0.740 | 0.856 | 0.872 | - | 0.556 | 0.724 |
199
  disambiguation_qa | 0.612 | 0.588 | 0.400 | 0.712 | 0.668 | 0.532 | 0.688 | 0.652 | 0.660 | 0.720 | 0.584 | 0.636 | 0.628 | 0.640 | 0.380 | 0.616 | 0.648 | 0.732 | 0.584 | 0.536 | 0.552 | 0.540 | 0.516 | 0.424 | 0.472 | 0.688 | 0.644 | 0.600 | 0.596 | 0.728 | 0.704 | 0.604 | 0.640 | 0.592 | 0.720 | 0.752 | 0.692 | 0.768 | 0.660 | 0.696 | 0.544 | 0.556 | 0.656 | 0.636 | 0.764 | 0.780 | - | 0.576 | 0.640 |
200
  dyck_languages | 0.596 | 0.664 | 0.424 | 0.704 | 0.712 | 0.476 | 0.752 | 0.720 | 0.728 | 0.600 | 0.516 | 0.544 | 0.560 | 0.704 | 0.356 | 0.700 | 0.664 | 0.728 | 0.792 | 0.228 | 0.832 | 0.724 | 0.796 | 0.536 | 0.680 | 0.736 | 0.756 | 0.744 | 0.712 | 0.664 | 0.732 | 0.752 | 0.532 | 0.424 | 0.580 | 0.468 | 0.532 | 0.776 | 0.756 | 0.576 | 0.596 | 0.628 | 0.868 | 0.836 | 0.648 | 0.820 | - | 0.684 | 0.572 |
 
243
  psychology | 0.502 | 0.642 | 0.258 | 0.621 | 0.645 | 0.494 | 0.666 | 0.657 | 0.647 | 0.692 | 0.586 | 0.602 | 0.436 | 0.531 | 0.184 | 0.448 | 0.543 | 0.572 | 0.560 | 0.581 | 0.624 | 0.601 | 0.637 | 0.317 | 0.518 | 0.604 | 0.595 | - | 0.588 | 0.543 | 0.552 | 0.626 | 0.573 | 0.583 | 0.621 | 0.572 | 0.676 | 0.759 | 0.595 | 0.536 | 0.526 | 0.563 | 0.636 | 0.644 | 0.721 | 0.749 | - | 0.525 | - |
244
  MMLUPRO | 0.416 | 0.554 | 0.186 | 0.552 | 0.517 | 0.326 | 0.524 | 0.553 | 0.528 | 0.568 | 0.471 | 0.480 | 0.298 | 0.395 | 0.145 | 0.324 | 0.410 | 0.432 | 0.404 | 0.475 | 0.494 | 0.491 | 0.499 | 0.215 | 0.419 | 0.539 | 0.458 | 0.453 | 0.436 | 0.376 | 0.382 | 0.475 | 0.405 | 0.451 | 0.482 | 0.408 | 0.470 | 0.719 | 0.475 | 0.368 | 0.430 | 0.457 | 0.564 | 0.575 | 0.649 | 0.671 | - | 0.326 | 0.509 |
245
  CATEGORIES
246
+ REASONING | 0.658 | 0.774 | 0.367 | 0.713 | 0.738 | 0.570 | 0.782 | 0.800 | 0.788 | 0.814 | 0.804 | 0.811 | 0.592 | 0.702 | 0.096 | 0.561 | 0.703 | 0.845 | 0.684 | 0.754 | 0.702 | 0.708 | 0.713 | 0.352 | 0.606 | 0.791 | 0.779 | 0.628 | 0.730 | 0.785 | 0.799 | 0.744 | 0.713 | 0.741 | 0.724 | 0.691 | 0.805 | 0.809 | 0.755 | 0.747 | 0.689 | 0.719 | 0.805 | 0.809 | 0.850 | 0.874 | 0.885 | 0.784 | 0.806 |
247
+ UNDERSTANDING | 0.548 | 0.652 | 0.366 | 0.644 | 0.670 | 0.538 | 0.708 | 0.712 | 0.707 | 0.742 | 0.661 | 0.670 | 0.511 | 0.598 | 0.116 | 0.481 | 0.599 | 0.685 | 0.602 | 0.680 | 0.622 | 0.618 | 0.631 | 0.330 | 0.579 | 0.672 | 0.633 | 0.563 | 0.633 | 0.644 | 0.651 | 0.661 | 0.629 | 0.629 | 0.614 | 0.622 | 0.727 | 0.728 | 0.674 | 0.649 | 0.605 | 0.613 | 0.692 | 0.696 | 0.761 | 0.793 | 0.809 | 0.617 | 0.713 |
248
  LANGUAGE | 0.613 | 0.680 | 0.524 | 0.688 | 0.692 | 0.624 | 0.733 | 0.732 | 0.735 | 0.755 | 0.786 | 0.783 | 0.746 | 0.799 | 0.665 | 0.732 | 0.790 | 0.732 | 0.710 | 0.729 | 0.740 | 0.738 | 0.747 | 0.610 | 0.705 | 0.715 | 0.776 | 0.766 | 0.714 | 0.744 | 0.733 | 0.613 | 0.638 | 0.618 | 0.677 | 0.613 | 0.632 | 0.750 | 0.735 | 0.564 | 0.685 | 0.682 | 0.722 | 0.724 | 0.769 | 0.781 | 0.780 | 0.654 | 0.708 |
249
+ KNOWLEDGE | 0.516 | 0.627 | 0.354 | 0.476 | 0.496 | 0.505 | 0.677 | 0.710 | 0.689 | 0.733 | 0.553 | 0.543 | 0.406 | 0.470 | 0.266 | 0.358 | 0.404 | 0.601 | 0.536 | 0.633 | 0.568 | 0.571 | 0.547 | 0.066 | 0.536 | 0.580 | 0.526 | 0.582 | 0.511 | 0.542 | 0.533 | 0.581 | 0.546 | 0.546 | 0.517 | 0.519 | 0.663 | 0.678 | 0.530 | 0.585 | 0.469 | 0.426 | 0.595 | 0.597 | 0.693 | 0.725 | 0.581 | 0.489 | 0.521 |
250
  COT | 0.438 | 0.561 | 0.220 | 0.552 | 0.530 | 0.350 | 0.548 | 0.568 | 0.550 | 0.582 | 0.485 | 0.500 | 0.336 | 0.431 | 0.188 | 0.365 | 0.445 | 0.505 | 0.446 | 0.488 | 0.522 | 0.521 | 0.530 | 0.255 | 0.446 | 0.545 | 0.479 | 0.498 | 0.466 | 0.416 | 0.424 | 0.502 | 0.437 | 0.474 | 0.506 | 0.440 | 0.503 | 0.725 | 0.492 | 0.398 | 0.443 | 0.462 | 0.570 | 0.581 | 0.653 | 0.684 | - | 0.377 | 0.563 |
251
+ MATHCOT | 0.731 | 0.819 | 0.369 | 0.745 | 0.752 | 0.480 | 0.728 | 0.733 | 0.735 | 0.740 | 0.682 | 0.679 | 0.545 | 0.638 | 0.359 | 0.571 | 0.649 | 0.700 | 0.622 | 0.680 | 0.708 | 0.721 | 0.728 | 0.386 | 0.647 | 0.808 | 0.750 | 0.493 | 0.652 | 0.546 | 0.578 | 0.695 | 0.612 | 0.591 | 0.767 | 0.638 | 0.665 | 0.919 | 0.671 | 0.548 | 0.667 | 0.694 | 0.823 | 0.829 | 0.869 | 0.903 | 0.927 | 0.535 | 0.662 |
252
+ CODE | 0.416 | 0.498 | 0.176 | 0.499 | 0.534 | 0.331 | 0.487 | 0.483 | 0.495 | 0.568 | 0.456 | 0.475 | 0.339 | 0.410 | 0.210 | 0.356 | 0.466 | 0.344 | 0.269 | 0.411 | 0.440 | 0.439 | 0.463 | 0.217 | 0.366 | 0.498 | 0.514 | 0.321 | 0.326 | 0.324 | 0.316 | 0.430 | 0.372 | 0.389 | 0.427 | 0.376 | 0.350 | 0.568 | 0.390 | 0.346 | 0.437 | 0.445 | 0.510 | 0.528 | 0.578 | 0.612 | 0.321 | 0.233 | 0.368 |
253
  DISCIPLINES
254
+ NLP | 0.620 | 0.720 | 0.408 | 0.647 | 0.670 | 0.568 | 0.751 | 0.767 | 0.755 | 0.786 | 0.729 | 0.728 | 0.588 | 0.667 | 0.263 | 0.549 | 0.648 | 0.774 | 0.650 | 0.713 | 0.675 | 0.678 | 0.677 | 0.329 | 0.609 | 0.722 | 0.723 | 0.642 | 0.685 | 0.739 | 0.737 | 0.681 | 0.655 | 0.667 | 0.647 | 0.637 | 0.744 | 0.755 | 0.693 | 0.687 | 0.632 | 0.630 | 0.731 | 0.734 | 0.791 | 0.818 | 0.772 | 0.712 | 0.725 |
255
+ MATH | 0.597 | 0.712 | 0.294 | 0.674 | 0.659 | 0.398 | 0.634 | 0.650 | 0.637 | 0.653 | 0.590 | 0.597 | 0.445 | 0.544 | 0.265 | 0.482 | 0.564 | 0.638 | 0.528 | 0.594 | 0.612 | 0.613 | 0.629 | 0.318 | 0.556 | 0.711 | 0.634 | 0.451 | 0.558 | 0.483 | 0.508 | 0.611 | 0.527 | 0.525 | 0.646 | 0.543 | 0.585 | 0.817 | 0.612 | 0.493 | 0.576 | 0.599 | 0.741 | 0.747 | 0.799 | 0.843 | 0.927 | 0.455 | 0.615 |
256
+ SCIENCE | 0.606 | 0.726 | 0.350 | 0.741 | 0.713 | 0.555 | 0.735 | 0.749 | 0.739 | 0.769 | 0.686 | 0.698 | 0.481 | 0.576 | 0.093 | 0.480 | 0.579 | 0.664 | 0.608 | 0.685 | 0.668 | 0.668 | 0.676 | 0.346 | 0.605 | 0.716 | 0.621 | 0.673 | 0.618 | 0.581 | 0.596 | 0.701 | 0.658 | 0.685 | 0.696 | 0.660 | 0.697 | 0.845 | 0.674 | 0.618 | 0.629 | 0.657 | 0.738 | 0.748 | 0.815 | 0.806 | 0.946 | 0.544 | 0.731 |
257
  ENGINEERING | 0.349 | 0.429 | 0.166 | 0.464 | 0.453 | 0.280 | 0.412 | 0.426 | 0.426 | 0.438 | 0.334 | 0.333 | 0.224 | 0.303 | 0.108 | 0.253 | 0.289 | 0.333 | 0.332 | 0.386 | 0.412 | 0.404 | 0.397 | 0.169 | 0.323 | 0.407 | 0.346 | 0.393 | 0.339 | 0.267 | 0.272 | 0.305 | 0.300 | 0.319 | 0.323 | 0.308 | 0.371 | 0.595 | 0.388 | 0.283 | 0.315 | 0.325 | 0.443 | 0.444 | 0.530 | 0.590 | - | 0.199 | 0.586 |
258
  MEDICINE | 0.379 | 0.504 | 0.216 | 0.524 | 0.540 | 0.400 | 0.591 | 0.590 | 0.595 | 0.648 | 0.521 | 0.530 | 0.346 | 0.464 | 0.069 | 0.347 | 0.469 | 0.515 | 0.544 | 0.620 | 0.568 | 0.572 | 0.577 | 0.243 | 0.496 | 0.543 | 0.512 | 0.447 | 0.541 | 0.485 | 0.503 | 0.558 | 0.507 | 0.525 | 0.537 | 0.501 | 0.633 | 0.672 | 0.510 | 0.482 | 0.459 | 0.478 | 0.574 | 0.580 | 0.655 | 0.702 | 0.598 | 0.457 | 0.635 |
259
+ HUMANITIES | 0.472 | 0.594 | 0.291 | 0.572 | 0.615 | 0.485 | 0.643 | 0.652 | 0.645 | 0.679 | 0.593 | 0.610 | 0.416 | 0.521 | 0.094 | 0.395 | 0.517 | 0.603 | 0.545 | 0.624 | 0.571 | 0.563 | 0.578 | 0.292 | 0.529 | 0.609 | 0.552 | 0.536 | 0.552 | 0.535 | 0.547 | 0.605 | 0.567 | 0.567 | 0.588 | 0.567 | 0.671 | 0.740 | 0.597 | 0.544 | 0.527 | 0.533 | 0.629 | 0.638 | 0.716 | 0.742 | 0.600 | 0.508 | 0.660 |
260
  BUSINESS | 0.536 | 0.657 | 0.252 | 0.679 | 0.655 | 0.450 | 0.660 | 0.697 | 0.678 | 0.709 | 0.623 | 0.637 | 0.408 | 0.523 | 0.115 | 0.428 | 0.530 | 0.594 | 0.537 | 0.626 | 0.592 | 0.582 | 0.598 | 0.251 | 0.517 | 0.667 | 0.564 | 0.466 | 0.565 | 0.514 | 0.526 | 0.632 | 0.575 | 0.589 | 0.637 | 0.604 | 0.636 | 0.801 | 0.644 | 0.566 | 0.565 | 0.596 | 0.701 | 0.710 | 0.759 | 0.802 | - | 0.479 | 0.652 |
261
+ LAW | 0.316 | 0.419 | 0.200 | 0.396 | 0.427 | 0.302 | 0.489 | 0.485 | 0.483 | 0.524 | 0.417 | 0.429 | 0.280 | 0.344 | 0.075 | 0.258 | 0.348 | 0.420 | 0.374 | 0.412 | 0.380 | 0.397 | 0.406 | 0.175 | 0.344 | 0.420 | 0.365 | 0.370 | 0.367 | 0.367 | 0.374 | 0.427 | 0.393 | 0.399 | 0.392 | 0.310 | 0.498 | 0.541 | 0.446 | 0.376 | 0.374 | 0.383 | 0.451 | 0.456 | 0.541 | 0.604 | - | 0.327 | 0.465 |
262
  COMPOSITE AVERAGE
263
+ AVG | 0.561 | 0.668 | 0.342 | 0.627 | 0.641 | 0.502 | 0.688 | 0.701 | 0.692 | 0.724 | 0.648 | 0.654 | 0.493 | 0.581 | 0.198 | 0.476 | 0.576 | 0.670 | 0.582 | 0.651 | 0.622 | 0.623 | 0.629 | 0.306 | 0.561 | 0.668 | 0.635 | 0.578 | 0.598 | 0.610 | 0.616 | 0.634 | 0.595 | 0.606 | 0.616 | 0.585 | 0.674 | 0.748 | 0.633 | 0.597 | 0.578 | 0.586 | 0.686 | 0.691 | 0.754 | 0.783 | 0.759 | 0.578 | 0.675 |
264
+
265
 
266
  CODE MODELS:
267