-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
1043 lines (1005 loc) · 30.2 KB
/
index.html
File metadata and controls
1043 lines (1005 loc) · 30.2 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<html lang="en">
<head>
<script src="https://cdn.jsdelivr.net/npm/chart.js"></script>
<meta charset="utf-8" />
<meta
name="description"
content="Benchmarking LLMs' Capabilities to Generate Structural Outputs"
/>
<meta
name="keywords"
content="StructEval, LMM, LMM Evaluation, Vision Language Model, Large Language Model, Large Multimodal Model, artificial intelligence, AI, AGI, artificial general intelligence"
/>
<meta name="viewport" content="width=device-width, initial-scale=1" />
<title>
StructEval: Benchmarking LLMs' Capabilities to Generate Structural Outputs
</title>
<link rel="icon" href="./static/images/imgs/logo.png" />
<link
href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet"
/>
<link rel="stylesheet" href="./static/css/bulma.min.css" />
<link rel="stylesheet" href="./static/css/bulma-carousel.min.css" />
<link rel="stylesheet" href="./static/css/bulma-slider.min.css" />
<link rel="stylesheet" href="./static/css/fontawesome.all.min.css" />
<link
rel="stylesheet"
href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css"
/>
<link rel="stylesheet" href="./static/css/index.css" />
<script
src="https://kit.fontawesome.com/fff5b27ec1.js"
crossorigin="anonymous"
></script>
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script defer src="./static/js/fontawesome.all.min.js"></script>
<script src="./static/js/bulma-carousel.min.js"></script>
<script src="./static/js/bulma-slider.min.js"></script>
<script src="./static/js/index.js"></script>
</head>
<body>
<nav class="navbar" role="navigation" aria-label="main navigation">
<div class="navbar-brand">
<a
role="button"
class="navbar-burger"
aria-label="menu"
aria-expanded="false"
>
<span aria-hidden="true"></span>
<span aria-hidden="true"></span>
<span aria-hidden="true"></span>
</a>
</div>
<div class="navbar-menu">
<div class="navbar-start" style="flex-grow: 1; justify-content: center">
<a class="navbar-item" href="#abstract">Abstract</a>
<a class="navbar-item" href="#dataset">Dataset</a>
<a class="navbar-item" href="#examples">Examples</a>
<a class="navbar-item" href="#evaluation">Evaluation</a>
<a class="navbar-item" href="#leaderboard">Leaderboard</a>
<a class="navbar-item" href="#analysis">Analysis</a>
<a class="navbar-item" href="#citation">Citation</a>
</div>
</div>
</nav>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title is-bold">
<span>
<img
src="static/images/figures/StructEvalLogo.png"
style="max-width: 150px; margin-bottom: 20px"
alt="StructEval Logo"
/>
</span>
<span>StructEval</span>
</h1>
<h2 class="subtitle is-3 publication-subtitle">
Benchmarking LLMs' Capabilities to Generate Structural Outputs
</h2>
<div
class="is-size-5 publication-authors"
style="
display: flex;
flex-direction: column;
align-items: center;
"
>
<p class="author-block">
Jialin Yang, Dongfu Jiang, Lipeng He, Sherman Siu, Yuxuan
Zhang, Disen Liao, Zhuofeng Li, Huaye Zeng, Yiming Jia, Haozhe
Wang, Benjamin Schneider, Chi Ruan, Wentao Ma, Zhiheng Lyu,
Yifei Wang, Yi Lu, Quy Duc Do, Ziyan Jiang, Ping Nie, Wenhu
Chen
</p>
<p style="margin-top: 25px; font-style: italic">
University of Waterloo
</p>
</div>
<br />
<div class="column has-text-centered">
<div class="publication-links">
<span class="link-block">
<a
href="#"
class="external-link button is-normal is-rounded is-dark"
>
<span class="icon">
<i class="fas fa-file-pdf"></i>
</span>
<span>Paper (PDF)</span>
</a>
</span>
<span class="link-block">
<a
href="https://github.com/TIGER-AI-Lab/StructEval"
class="external-link button is-normal is-rounded is-dark"
>
<span class="icon">
<i class="fab fa-github"></i>
</span>
<span>Code (GitHub)</span>
</a>
</span>
<span class="link-block">
<a
href="#leaderboard"
class="external-link button is-normal is-rounded is-dark"
>
<span class="icon has-text-white">
<i class="fa-solid fa-trophy"></i>
</span>
<span>Leaderboard</span>
</a>
</span>
<span class="link-block">
<a
href="https://huggingface.co/datasets/TIGER-Lab/StructEval"
class="external-link button is-normal is-rounded is-dark"
>
<span class="icon has-text-white">🤗</span>
<span>Hugging Face</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="hero teaser" style="padding: 50px 0px">
<div class="container is-max-desktop has-text-centered">
<img
src="static/images/figures/intro.png"
alt="StructEval Overview"
style="max-width: 75%; margin-left: 48px"
/>
<p style="margin-top: 15px">
Our benchmark encompasses 18 distinct formats and 44 task types
organized into two complementary subsets: <i>StructEval-T</i>, which
assesses the generation of text-only structures such as JSON and TOML,
and <i>StructEval-V</i>, which evaluates the quality of visually
rendered outputs from code such as HTML and SVG.
</p>
</div>
</section>
<section class="section" id="abstract" style="padding-top: 70px">
<div class="container is-max-desktop">
<h2 class="title is-3 has-text-centered" style="margin-bottom: 25px">
Abstract
</h2>
<div class="content has-text-justified">
<p>
As Large Language Models (LLMs) become integral to software
development workflows, their ability to generate structured outputs
has become critically important. We introduce <b>StructEval</b>, a
comprehensive benchmark for evaluating LLMs' capabilities in
producing both non-renderable (JSON, YAML, CSV) and renderable
(HTML, React, SVG) structured formats. Unlike prior benchmarks,
StructEval systematically evaluates structural fidelity across
diverse formats through two paradigms: (1) generation tasks,
producing structured output from natural language prompts, and (2)
conversion tasks, translating between structured formats.
</p>
<p>
Our benchmark encompasses 18 formats and 44 types of tasks, with
novel metrics for format adherence and structural correctness.
Results reveal significant performance gaps—even state-of-the-art
models like o1-mini achieve only 75.58% average score, with
open-source alternatives lagging approximately 10 points behind. We
find generation tasks more challenging than conversion tasks, and
producing correct visual content more difficult than generating
text-only structures.
</p>
</div>
</div>
</section>
<hr />
<section class="section" id="dataset" style="padding-top: 50px">
<div class="container is-max-desktop">
<h2 class="title is-3 has-text-centered" style="margin-bottom: 25px">
Dataset Overview
</h2>
<div class="content has-text-justified">
<p>
<b>StructEval</b> comprises <b>2,035 examples</b> covering
<b>44 unique structure generation tasks</b> across
<b>18 structured output formats</b>. The dataset is organized into
two main subsets:
</p>
<div style="text-align: center; width: 100%; margin-top: 30px">
<div
class="box"
style="
background: #fafafa;
padding: 30px;
border-radius: 8px;
margin: 20px auto;
width: 100%;
box-sizing: border-box;
"
>
<h5 style="font-size: 1.25rem; margin-bottom: 25px">
All 18 Supported Formats
</h5>
<div
style="
display: grid;
grid-template-columns: repeat(auto-fit, minmax(120px, 1fr));
width: 100%;
border: 0.5px solid #ccc;
gap: 0;
"
>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
JSON
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
CSV
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
XML
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
YAML
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
TOML
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
HTML
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
React
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
Vue
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
Angular
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
SVG
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
LaTeX
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
Markdown
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
Mermaid
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
TikZ
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
Matplotlib
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
Canvas
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
Vega
</div>
<div
style="
border: 0.5px solid #ccc;
padding: 20px;
text-align: center;
"
>
Typst
</div>
</div>
</div>
</div>
<div class="columns is-centered" style="margin-top: 30px">
<div class="column half">
<div class="box has-background-light">
<h4
class="title is-5 has-text-centered"
style="margin-bottom: 10px"
>
📝 StructEval-T
</h4>
<p class="has-text-centered">
Evaluates text-only structured outputs
</p>
<hr
style="
background-color: rgb(122 122 122 / 40%);
margin-top: 15px;
"
/>
<ul>
<li><b>Formats:</b> JSON, XML, YAML, CSV, TOML</li>
<li><b>Tasks:</b> 19 (5 generation, 14 conversion)</li>
<li><b>Examples:</b> 950</li>
<li>
<b>Focus:</b> Syntactic validity & structural correctness
</li>
</ul>
</div>
</div>
<div class="column half">
<div class="box has-background-info-light">
<h4 class="title has-text-centered" style="margin-bottom: 10px">
🎨 StructEval-V
</h4>
<p class="has-text-centered">
Evaluates visually rendered outputs
</p>
<hr
style="
background-color: rgb(122 122 122 / 40%);
margin-top: 15px;
"
/>
<ul>
<li>
<b>Formats:</b> HTML, React, SVG, LaTeX, Mermaid, etc.
</li>
<li><b>Tasks:</b> 25 (13 generation, 12 conversion)</li>
<li><b>Examples:</b> 1,085</li>
<li><b>Focus:</b> Visual correctness via VQA evaluation</li>
</ul>
</div>
</div>
</div>
</div>
</div>
</section>
<hr />
<!-- Examples Section -->
<section class="section" id="examples" style="padding: 25px 0px">
<div class="container">
<h2 class="title is-3 has-text-centered">Example Tasks</h2>
<div class="columns is-centered">
<div class="column is-9">
<div class="tabs is-boxed is-centered">
<ul>
<li class="is-active">
<a onclick="showExample('generation')">
<span>Generation Tasks</span>
</a>
</li>
<li>
<a onclick="showExample('conversion')">
<span>Conversion Tasks</span>
</a>
</li>
<li>
<a onclick="showExample('rendered')">
<span>Rendered Outputs</span>
</a>
</li>
</ul>
</div>
<div id="generation-examples" class="example-content">
<div class="box">
<h4 class="title is-5">📝 StructEval-T Generation Example</h4>
<div class="columns">
<div class="column is-6">
<h6 class="subtitle is-6">Task Prompt:</h6>
<pre style="font-size: 0.85em">
Please output JSON code.
Task:
Summarize metadata about a fictional
scientific article.
Feature Requirements:
1. Top-level field "title" is a string
2. Field "authors" is a list of exactly two items
3. Each author has "name" and "affiliation"
4. Field "publication.year" is an integer
5. Field "keywords" is a list of strings</pre
>
</div>
<div class="column is-6">
<h6 class="subtitle is-6">Expected Keywords:</h6>
<ul>
<li><code>title</code></li>
<li><code>authors[0].name</code></li>
<li><code>authors[1].affiliation</code></li>
<li><code>publication.year</code></li>
<li><code>keywords[2]</code></li>
</ul>
</div>
</div>
</div>
<div class="box">
<h4 class="title is-5">🎨 StructEval-V Generation Example</h4>
<div class="columns">
<div class="column is-6">
<h6 class="subtitle is-6">Task Prompt:</h6>
<pre style="font-size: 0.85em">
Please output HTML code.
Task:
Design a webpage for a travel itinerary.
Feature Requirements:
• Centered <h1> with "Trip Summary"
• Use a <table> with 3 rows and 2 columns
• Apply class "highlight" to second row
• Add <button> labeled "Export PDF"</pre
>
</div>
<div class="column is-6">
<h6 class="subtitle is-6">VQA Evaluation:</h6>
<ul>
<li>
<b>Q:</b> What text is in the h1 header?<br /><b>A:</b>
Trip Summary
</li>
<li>
<b>Q:</b> How many rows are in the table?<br /><b>A:</b>
3
</li>
<li>
<b>Q:</b> What class is on the second row?<br /><b
>A:</b
>
highlight
</li>
<li>
<b>Q:</b> What text is on the button?<br /><b>A:</b>
Export PDF
</li>
</ul>
</div>
</div>
</div>
</div>
<div
id="conversion-examples"
class="example-content"
style="display: none"
>
<div class="box">
<h4 class="title is-5">Format Conversion Examples</h4>
<div class="columns">
<div class="column is-6">
<h6 class="subtitle is-6">JSON → YAML</h6>
<pre style="font-size: 0.85em">
{
"name": "StructEval",
"version": "1.0",
"tasks": ["generation", "conversion"],
"formats": {
"text": ["JSON", "YAML", "XML"],
"visual": ["HTML", "React", "SVG"]
}
}</pre
>
</div>
<div class="column is-6">
<h6 class="subtitle is-6">Expected YAML Output</h6>
<pre style="font-size: 0.85em">
name: StructEval
version: "1.0"
tasks:
- generation
- conversion
formats:
text:
- JSON
- YAML
- XML
visual:
- HTML
- React
- SVG</pre
>
</div>
</div>
</div>
</div>
<div
id="rendered-examples"
class="example-content"
style="display: none"
>
<h4 class="title is-5 has-text-centered">
Example Rendered Outputs
</h4>
<div class="columns is-multiline">
<div class="column is-4">
<div class="box">
<figure class="image">
<img src="static/images/imgs/000107.png" alt="Angular" />
</figure>
<p class="has-text-centered" style="margin-top: 15px">
<b>Angular</b>
</p>
</div>
</div>
<div class="column is-4">
<div class="box">
<figure class="image">
<img
src="static/images/imgs/000829.png"
alt="Matplotlib"
/>
</figure>
<p class="has-text-centered" style="margin-top: 15px">
<b>Matplotlib</b>
</p>
</div>
</div>
<div class="column is-4">
<div class="box">
<figure class="image">
<img src="static/images/imgs/000936.png" alt="Mermaid" />
</figure>
<p class="has-text-centered" style="margin-top: 15px">
<b>Mermaid</b>
</p>
</div>
</div>
<div class="column is-4">
<div class="box">
<figure class="image">
<img src="static/images/imgs/001126.png" alt="React" />
</figure>
<p class="has-text-centered" style="margin-top: 15px">
<b>React</b>
</p>
</div>
</div>
<div class="column is-4">
<div class="box">
<figure class="image">
<img src="static/images/imgs/001210.png" alt="SVG" />
</figure>
<p class="has-text-centered" style="margin-top: 15px">
<b>SVG</b>
</p>
</div>
</div>
<div class="column is-4">
<div class="box">
<figure class="image">
<img src="static/images/imgs/001312.png" alt="TikZ" />
</figure>
<p class="has-text-centered" style="margin-top: 15px">
<b>TikZ</b>
</p>
</div>
</div>
<div class="column is-4">
<div class="box">
<figure class="image">
<img src="static/images/imgs/001509.png" alt="Vega" />
</figure>
<p class="has-text-centered" style="margin-top: 15px">
<b>Vega</b>
</p>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<hr />
<section class="section" id="evaluation">
<div class="container">
<h2 class="title is-3 has-text-centered">Evaluation Framework</h2>
<div class="columns is-centered">
<div class="column is-9">
<div class="content has-text-centered" style="margin-bottom: 50px">
<img
src="static/images/imgs/annotation_pipeline.png"
alt="Evaluation Pipeline"
style="max-width: 90%; margin: 20px auto"
/>
</div>
<div class="content has-text-justified">
<p>Our evaluation framework employs four core metrics:</p>
<div class="columns" style="margin-top: 10px">
<div class="column is-3">
<div class="box">
<h5 class="title is-5">🔧 Render Score (T/V)</h5>
<p style="text-align: left">
Measures the percentage of generated code that can be
successfully loaded or rendered into images
</p>
</div>
</div>
<div class="column is-3">
<div class="box">
<h5 class="title is-5">✓ Syntax Score (T)</h5>
<p style="text-align: left">
Verifies structural correctness using format-specific
parsers or rendering engines
</p>
</div>
</div>
<div class="column is-3">
<div class="box">
<h5 class="title is-5">🔍 Keyword Matching (V)</h5>
<p style="text-align: left">
Evaluates presence of required structural elements using
path-based rules
</p>
</div>
</div>
<div class="column is-3">
<div class="box">
<h5 class="title is-5">👁️ VQA Score (V)</h5>
<p style="text-align: left">
Assesses visual correctness through GPT-4V-based question
answering (V tasks only)
</p>
</div>
</div>
</div>
<div style="margin-top: 40px">
<h3 class="title is-4 has-text-centered">Score Aggregation Methods</h3>
<div class="columns" style="margin-top: 20px">
<div class="column is-6">
<div class="box has-background-info-light">
<h5 class="title is-5 has-text-centered">🎨 Renderable Formats</h5>
<p class="has-text-centered" style="margin-bottom: 15px">
<em>HTML, React, SVG, LaTeX, Mermaid, etc.</em>
</p>
<div class="content">
<pre style="background: #f5f5f5; padding: 15px; border-radius: 5px; font-size: 0.9em;">final_score = (0.2 × render_score) +
(0.1 × keyword_matching) +
(0.7 × vqa_score)</pre>
</div>
</div>
</div>
<div class="column is-6">
<div class="box has-background-light">
<h5 class="title is-5 has-text-centered">📝 Non-Renderable Formats</h5>
<p class="has-text-centered" style="margin-bottom: 15px">
<em>JSON, XML, YAML, CSV, TOML</em>
</p>
<div class="content">
<pre style="background: #f5f5f5; padding: 15px; border-radius: 5px; font-size: 0.9em;">final_score = (0.2 × render_score) +
(0.8 × syntax_score)</pre>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<hr />
<section
class="section"
style="padding-bottom: 20px; padding-top: 0px"
id="leaderboard"
>
<div class="container">
<div class="columns is-centered">
<div class="column is-full has-text-centered content is-9">
<h2 class="title is-3" id="leaderboard">Leaderboard</h2>
<div class="content has-text-justified">
<p>
We evaluate various state-of-the-art LLMs in a zero-shot
setting. The table below shows the performance breakdown across
our four task categories. Click on the column headers to expand
detailed results.
</p>
</div>
<div class="model-labels-container">
<span class="leaderboard-label open_source">Open-Source</span>
<span class="leaderboard-label proprietary">Closed-Source</span>
</div>
<br />
<div class="leaderboard-container">
<div class="table-wrapper">
<table id="structeval-table">
<thead>
<tr>
<th
rowspan="2"
class="sortable clickable"
data-sort="string"
>
Model
</th>
<th
rowspan="2"
class="sortable clickable"
data-sort="string"
>
Type
</th>
<th colspan="2" style="text-align: center">
StructEval-T
</th>
<th colspan="2" style="text-align: center">
StructEval-V
</th>
<th
rowspan="2"
class="sortable clickable"
data-sort="number"
>
Average
</th>
</tr>
<tr>
<th class="sortable clickable" data-sort="number">
Generation
</th>
<th class="sortable clickable" data-sort="number">
Conversion
</th>
<th class="sortable clickable" data-sort="number">
Generation
</th>
<th class="sortable clickable" data-sort="number">
Conversion
</th>
</tr>
</thead>
<tbody>
<!-- Table body will be populated dynamically -->
</tbody>
</table>
</div>
</div>
</div>
</div>
<hr />
<div class="columns is-centered" id="analysis">
<div class="column is-full has-text-centered content is-9">
<h2 class="title is-3">Performance Analysis</h2>
<div class="columns" style="margin-top: 25px">
<div class="column is-6">
<canvas id="task_type_chart"></canvas>
<p class="has-text-centered">Performance by Task Type</p>
</div>
<div class="column is-6">
<canvas id="format_difficulty_chart"></canvas>
<p class="has-text-centered">
Challenging Formats (Average Score < 50%)
</p>
</div>
</div>
</div>
</div>
<hr />
<div class="container" style="padding-top: 25px; padding-bottom: 50px">
<h2 class="title is-3 has-text-centered">Key Findings</h2>
<div class="columns is-centered" style="margin-top: 25px">
<div class="column is-9">
<div class="columns is-multiline">
<div class="column is-6">
<div class="box" style="height: 100%">
<h5 class="title is-5">
<span class="icon has-text-danger">
<i class="fas fa-chart-line"></i>
</span>
Performance Gap
</h5>
<p>
Even state-of-the-art models struggle with structured
output generation. GPT-4o achieves only
<b>76.02%</b> average score, while the best open-source
model (Qwen3-4B) lags at <b>67.04%</b>.
</p>
</div>
</div>
<div class="column is-6">
<div class="box" style="height: 100%">
<h5 class="title is-5">
<span class="icon has-text-info">
<i class="fas fa-balance-scale"></i>
</span>
Task Difficulty
</h5>
<p>
Generation tasks are generally
<b>more challenging</b> than conversion tasks. Visual
rendering (StructEval-V) proves <b>harder</b> than
text-only structures (StructEval-T).
</p>
</div>
</div>
<div class="column is-6">
<div class="box" style="height: 100%">
<h5 class="title is-5">
<span class="icon has-text-warning">
<i class="fas fa-exclamation-triangle"></i>
</span>
Challenging Formats
</h5>
<p>
Several formats remain particularly difficult for all
models: <b>Text→TOML</b> (35.8%),
<b>Text→Mermaid</b> (18.9%), and
<b>Matplotlib→TikZ</b> (28.4%) conversions.
</p>
</div>
</div>
<div class="column is-6">
<div class="box" style="height: 100%">
<h5 class="title is-5">
<span class="icon has-text-success">
<i class="fas fa-check-circle"></i>
</span>
Saturated Tasks
</h5>
<p>
Some tasks are effectively solved with scores >90%: JSON,
HTML, CSV generation and YAML→JSON, React→HTML conversions
show near-perfect performance.
</p>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<hr />
<section
class="section"
id="citation"
style="padding-top: 25px; padding-bottom: 100px"
>
<div class="container is-max-desktop content">
<h2 class="title is-3 has-text-centered">Citation</h2>
<pre><code>@inproceedings{yue2023mmmu,
title={MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI},
author={Xiang Yue and Yuansheng Ni and Kai Zhang and Tianyu Zheng and Ruoqi Liu and Ge Zhang and Samuel Stevens and Dongfu Jiang and Weiming Ren and Yuxuan Sun and Cong Wei and Botao Yu and Ruibin Yuan and Renliang Sun and Ming Yin and Boyuan Zheng and Zhenzhu Yang and Yibo Liu and Wenhao Huang and Huan Sun and Yu Su and Wenhu Chen},
booktitle={Proceedings of CVPR},
year={2024},
}</code></pre>