Hugging Face's logo

Datasets:
Ydsun
/
SynBench

Tasks:
Text Classification
Modalities:
Text
Formats:
json
Languages:
English
Size:
1M - 10M
Tags:
Synthetic
differential-privacy
membership-inference
benchmark
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
SynBench
File size: 35,101 Bytes
df5f28c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bcc785a
 
df5f28c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: other
license_name: per-subset
license_link: LICENSE.md
pretty_name: "SynBench: A Benchmark for Differentially Private Text Generation"
language:
- en
task_categories:
- text-classification
tags:
- synthetic
- differential-privacy
- membership-inference
- benchmark
size_categories:
- 100K<n<1M
configs:
- config_name: hallmarks_of_cancer
  data_files:
  - split: train
    path: hallmarks_of_cancer/original/train-original.jsonl
  - split: test
    path: hallmarks_of_cancer/original/test-original.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_0_5
    path: hallmarks_of_cancer/Llama-3.2-1B/dp-transformers/train0.5-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_1
    path: hallmarks_of_cancer/Llama-3.2-1B/dp-transformers/train1-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_2
    path: hallmarks_of_cancer/Llama-3.2-1B/dp-transformers/train2-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_4
    path: hallmarks_of_cancer/Llama-3.2-1B/dp-transformers/train4-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_inf
    path: hallmarks_of_cancer/Llama-3.2-1B/dp-transformers/trainnodp-dp-transformers.jsonl
  - split: llama_3_2_1b__aug_pe__eps_0_5
    path: hallmarks_of_cancer/Llama-3.2-1B/aug-pe/train0.5-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_1
    path: hallmarks_of_cancer/Llama-3.2-1B/aug-pe/train1-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_2
    path: hallmarks_of_cancer/Llama-3.2-1B/aug-pe/train2-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_4
    path: hallmarks_of_cancer/Llama-3.2-1B/aug-pe/train4-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_inf
    path: hallmarks_of_cancer/Llama-3.2-1B/aug-pe/trainnodp-aug-pe.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_0_5
    path: hallmarks_of_cancer/Llama-3.2-1B/dp-diffusion/train0.5-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_1
    path: hallmarks_of_cancer/Llama-3.2-1B/dp-diffusion/train1-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_2
    path: hallmarks_of_cancer/Llama-3.2-1B/dp-diffusion/train2-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_4
    path: hallmarks_of_cancer/Llama-3.2-1B/dp-diffusion/train4-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_inf
    path: hallmarks_of_cancer/Llama-3.2-1B/dp-diffusion/trainnodp-dp-diffusion.jsonl
  - split: llama_3_2_1b__epsvec__eps_0_5
    path: hallmarks_of_cancer/Llama-3.2-1B/EPSVec/train0.5-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_1
    path: hallmarks_of_cancer/Llama-3.2-1B/EPSVec/train1-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_2
    path: hallmarks_of_cancer/Llama-3.2-1B/EPSVec/train2-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_4
    path: hallmarks_of_cancer/Llama-3.2-1B/EPSVec/train4-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_inf
    path: hallmarks_of_cancer/Llama-3.2-1B/EPSVec/trainnodp-EPSVec.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_inf
    path: hallmarks_of_cancer/OLMo-2-0425-1B-10epochs/dp-transformers/train0-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_0_5
    path: hallmarks_of_cancer/OLMo-2-0425-1B-10epochs/dp-transformers/train0.5-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_1
    path: hallmarks_of_cancer/OLMo-2-0425-1B-10epochs/dp-transformers/train1-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_2
    path: hallmarks_of_cancer/OLMo-2-0425-1B-10epochs/dp-transformers/train2-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_4
    path: hallmarks_of_cancer/OLMo-2-0425-1B-10epochs/dp-transformers/train4-dp-transformers.jsonl
- config_name: DMSAFN
  data_files:
  - split: train
    path: DMSAFN/original/train-original.jsonl
  - split: test
    path: DMSAFN/original/test-original.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_0_5
    path: DMSAFN/Llama-3.2-1B/dp-transformers/train0.5-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_1
    path: DMSAFN/Llama-3.2-1B/dp-transformers/train1-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_2
    path: DMSAFN/Llama-3.2-1B/dp-transformers/train2-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_4
    path: DMSAFN/Llama-3.2-1B/dp-transformers/train4-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_inf
    path: DMSAFN/Llama-3.2-1B/dp-transformers/trainnodp-dp-transformers.jsonl
  - split: llama_3_2_1b__aug_pe__eps_0_5
    path: DMSAFN/Llama-3.2-1B/aug-pe/train0.5-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_1
    path: DMSAFN/Llama-3.2-1B/aug-pe/train1-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_2
    path: DMSAFN/Llama-3.2-1B/aug-pe/train2-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_4
    path: DMSAFN/Llama-3.2-1B/aug-pe/train4-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_inf
    path: DMSAFN/Llama-3.2-1B/aug-pe/trainnodp-aug-pe.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_0_5
    path: DMSAFN/Llama-3.2-1B/dp-diffusion/train0.5-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_1
    path: DMSAFN/Llama-3.2-1B/dp-diffusion/train1-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_2
    path: DMSAFN/Llama-3.2-1B/dp-diffusion/train2-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_4
    path: DMSAFN/Llama-3.2-1B/dp-diffusion/train4-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_inf
    path: DMSAFN/Llama-3.2-1B/dp-diffusion/trainnodp-dp-diffusion.jsonl
  - split: llama_3_2_1b__epsvec__eps_0_5
    path: DMSAFN/Llama-3.2-1B/EPSVec/train0.5-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_1
    path: DMSAFN/Llama-3.2-1B/EPSVec/train1-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_2
    path: DMSAFN/Llama-3.2-1B/EPSVec/train2-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_4
    path: DMSAFN/Llama-3.2-1B/EPSVec/train4-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_inf
    path: DMSAFN/Llama-3.2-1B/EPSVec/trainnodp-EPSVec.jsonl
  - split: llama_3_2_1b_instruct__dp_transformers__eps_0_5
    path: DMSAFN/Llama-3.2-1B-Instruct/dp-transformers/train0.5-dp-transformers-Llama-3.2-1B-Instruct.jsonl
  - split: llama_3_2_1b_instruct__dp_transformers__eps_1
    path: DMSAFN/Llama-3.2-1B-Instruct/dp-transformers/train1-dp-transformers-Llama-3.2-1B-Instruct.jsonl
  - split: llama_3_2_1b_instruct__dp_transformers__eps_2
    path: DMSAFN/Llama-3.2-1B-Instruct/dp-transformers/train2-dp-transformers-Llama-3.2-1B-Instruct.jsonl
  - split: llama_3_2_1b_instruct__dp_transformers__eps_4
    path: DMSAFN/Llama-3.2-1B-Instruct/dp-transformers/train4-dp-transformers-Llama-3.2-1B-Instruct.jsonl
  - split: llama_3_2_1b_instruct__dp_transformers__eps_inf
    path: DMSAFN/Llama-3.2-1B-Instruct/dp-transformers/trainnodp-dp-transformers-Llama-3.2-1B-Instruct.jsonl
  - split: llama_3_2_3b_instruct__dp_transformers__eps_0_5
    path: DMSAFN/Llama-3.2-3B-Instruct/dp-transformers/train0.5-dp-transformers-Llama-3.2-3B-Instruct.jsonl
  - split: llama_3_2_3b_instruct__dp_transformers__eps_1
    path: DMSAFN/Llama-3.2-3B-Instruct/dp-transformers/train1-dp-transformers-Llama-3.2-3B-Instruct.jsonl
  - split: llama_3_2_3b_instruct__dp_transformers__eps_2
    path: DMSAFN/Llama-3.2-3B-Instruct/dp-transformers/train2-dp-transformers-Llama-3.2-3B-Instruct.jsonl
  - split: llama_3_2_3b_instruct__dp_transformers__eps_4
    path: DMSAFN/Llama-3.2-3B-Instruct/dp-transformers/train4-dp-transformers-Llama-3.2-3B-Instruct.jsonl
  - split: llama_3_2_3b_instruct__dp_transformers__eps_inf
    path: DMSAFN/Llama-3.2-3B-Instruct/dp-transformers/trainnodp-dp-transformers-Llama-3.2-3B-Instruct.jsonl
  - split: llama_3_1_8b_instruct__dp_transformers__eps_0_5
    path: DMSAFN/Llama-3.1-8B-Instruct/dp-transformers/train0.5-dp-transformers-Llama-3.1-8B-Instruct.jsonl
  - split: llama_3_1_8b_instruct__dp_transformers__eps_1
    path: DMSAFN/Llama-3.1-8B-Instruct/dp-transformers/train1-dp-transformers-Llama-3.1-8B-Instruct.jsonl
  - split: llama_3_1_8b_instruct__dp_transformers__eps_2
    path: DMSAFN/Llama-3.1-8B-Instruct/dp-transformers/train2-dp-transformers-Llama-3.1-8B-Instruct.jsonl
  - split: llama_3_1_8b_instruct__dp_transformers__eps_4
    path: DMSAFN/Llama-3.1-8B-Instruct/dp-transformers/train4-dp-transformers-Llama-3.1-8B-Instruct.jsonl
  - split: llama_3_1_8b_instruct__dp_transformers__eps_inf
    path: DMSAFN/Llama-3.1-8B-Instruct/dp-transformers/trainnodp-dp-transformers-Llama-3.1-8B-Instruct.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_inf
    path: DMSAFN/OLMo-2-0425-1B-10epochs/dp-transformers/train0-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_0_5
    path: DMSAFN/OLMo-2-0425-1B-10epochs/dp-transformers/train0.5-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_1
    path: DMSAFN/OLMo-2-0425-1B-10epochs/dp-transformers/train1-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_2
    path: DMSAFN/OLMo-2-0425-1B-10epochs/dp-transformers/train2-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_4
    path: DMSAFN/OLMo-2-0425-1B-10epochs/dp-transformers/train4-dp-transformers.jsonl
- config_name: Lc37Fnd
  data_files:
  - split: train
    path: Lc37Fnd/original/train-original.jsonl
  - split: test
    path: Lc37Fnd/original/test-original.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_0_5
    path: Lc37Fnd/Llama-3.2-1B/dp-transformers/train0.5-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_1
    path: Lc37Fnd/Llama-3.2-1B/dp-transformers/train1-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_2
    path: Lc37Fnd/Llama-3.2-1B/dp-transformers/train2-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_4
    path: Lc37Fnd/Llama-3.2-1B/dp-transformers/train4-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_inf
    path: Lc37Fnd/Llama-3.2-1B/dp-transformers/trainnodp-dp-transformers.jsonl
  - split: llama_3_2_1b__aug_pe__eps_0_5
    path: Lc37Fnd/Llama-3.2-1B/aug-pe/train0.5-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_1
    path: Lc37Fnd/Llama-3.2-1B/aug-pe/train1-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_2
    path: Lc37Fnd/Llama-3.2-1B/aug-pe/train2-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_4
    path: Lc37Fnd/Llama-3.2-1B/aug-pe/train4-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_inf
    path: Lc37Fnd/Llama-3.2-1B/aug-pe/trainnodp-aug-pe.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_0_5
    path: Lc37Fnd/Llama-3.2-1B/dp-diffusion/train0.5-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_1
    path: Lc37Fnd/Llama-3.2-1B/dp-diffusion/train1-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_2
    path: Lc37Fnd/Llama-3.2-1B/dp-diffusion/train2-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_4
    path: Lc37Fnd/Llama-3.2-1B/dp-diffusion/train4-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_inf
    path: Lc37Fnd/Llama-3.2-1B/dp-diffusion/trainnodp-dp-diffusion.jsonl
  - split: llama_3_2_1b__epsvec__eps_0_5
    path: Lc37Fnd/Llama-3.2-1B/EPSVec/train0.5-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_1
    path: Lc37Fnd/Llama-3.2-1B/EPSVec/train1-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_2
    path: Lc37Fnd/Llama-3.2-1B/EPSVec/train2-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_4
    path: Lc37Fnd/Llama-3.2-1B/EPSVec/train4-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_inf
    path: Lc37Fnd/Llama-3.2-1B/EPSVec/trainnodp-EPSVec.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_inf
    path: Lc37Fnd/OLMo-2-0425-1B-10epochs/dp-transformers/train0-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_0_5
    path: Lc37Fnd/OLMo-2-0425-1B-10epochs/dp-transformers/train0.5-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_1
    path: Lc37Fnd/OLMo-2-0425-1B-10epochs/dp-transformers/train1-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_2
    path: Lc37Fnd/OLMo-2-0425-1B-10epochs/dp-transformers/train2-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_4
    path: Lc37Fnd/OLMo-2-0425-1B-10epochs/dp-transformers/train4-dp-transformers.jsonl
- config_name: ZSTFNS
  data_files:
  - split: train
    path: ZSTFNS/original/train-original.jsonl
  - split: test
    path: ZSTFNS/original/test-original.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_0_5
    path: ZSTFNS/Llama-3.2-1B/dp-transformers/train0.5-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_1
    path: ZSTFNS/Llama-3.2-1B/dp-transformers/train1-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_2
    path: ZSTFNS/Llama-3.2-1B/dp-transformers/train2-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_4
    path: ZSTFNS/Llama-3.2-1B/dp-transformers/train4-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_inf
    path: ZSTFNS/Llama-3.2-1B/dp-transformers/trainnodp-dp-transformers.jsonl
  - split: llama_3_2_1b__aug_pe__eps_0_5
    path: ZSTFNS/Llama-3.2-1B/aug-pe/train0.5-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_1
    path: ZSTFNS/Llama-3.2-1B/aug-pe/train1-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_2
    path: ZSTFNS/Llama-3.2-1B/aug-pe/train2-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_4
    path: ZSTFNS/Llama-3.2-1B/aug-pe/train4-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_inf
    path: ZSTFNS/Llama-3.2-1B/aug-pe/trainnodp-aug-pe.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_0_5
    path: ZSTFNS/Llama-3.2-1B/dp-diffusion/train0.5-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_1
    path: ZSTFNS/Llama-3.2-1B/dp-diffusion/train1-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_2
    path: ZSTFNS/Llama-3.2-1B/dp-diffusion/train2-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_4
    path: ZSTFNS/Llama-3.2-1B/dp-diffusion/train4-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_inf
    path: ZSTFNS/Llama-3.2-1B/dp-diffusion/trainnodp-dp-diffusion.jsonl
  - split: llama_3_2_1b__epsvec__eps_0_5
    path: ZSTFNS/Llama-3.2-1B/EPSVec/train0.5-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_1
    path: ZSTFNS/Llama-3.2-1B/EPSVec/train1-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_2
    path: ZSTFNS/Llama-3.2-1B/EPSVec/train2-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_4
    path: ZSTFNS/Llama-3.2-1B/EPSVec/train4-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_inf
    path: ZSTFNS/Llama-3.2-1B/EPSVec/trainnodp-EPSVec.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_inf
    path: ZSTFNS/OLMo-2-0425-1B-10epochs/dp-transformers/train0-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_0_5
    path: ZSTFNS/OLMo-2-0425-1B-10epochs/dp-transformers/train0.5-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_1
    path: ZSTFNS/OLMo-2-0425-1B-10epochs/dp-transformers/train1-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_2
    path: ZSTFNS/OLMo-2-0425-1B-10epochs/dp-transformers/train2-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_4
    path: ZSTFNS/OLMo-2-0425-1B-10epochs/dp-transformers/train4-dp-transformers.jsonl
- config_name: AYSLAX
  data_files:
  - split: train
    path: AYSLAX/original/train-original.jsonl
  - split: test
    path: AYSLAX/original/test-original.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_0_5
    path: AYSLAX/Llama-3.2-1B/dp-transformers/train0.5-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_1
    path: AYSLAX/Llama-3.2-1B/dp-transformers/train1-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_2
    path: AYSLAX/Llama-3.2-1B/dp-transformers/train2-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_4
    path: AYSLAX/Llama-3.2-1B/dp-transformers/train4-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_inf
    path: AYSLAX/Llama-3.2-1B/dp-transformers/trainnodp-dp-transformers.jsonl
  - split: llama_3_2_1b__aug_pe__eps_0_5
    path: AYSLAX/Llama-3.2-1B/aug-pe/train0.5-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_1
    path: AYSLAX/Llama-3.2-1B/aug-pe/train1-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_2
    path: AYSLAX/Llama-3.2-1B/aug-pe/train2-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_4
    path: AYSLAX/Llama-3.2-1B/aug-pe/train4-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_inf
    path: AYSLAX/Llama-3.2-1B/aug-pe/trainnodp-aug-pe.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_0_5
    path: AYSLAX/Llama-3.2-1B/dp-diffusion/train0.5-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_1
    path: AYSLAX/Llama-3.2-1B/dp-diffusion/train1-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_2
    path: AYSLAX/Llama-3.2-1B/dp-diffusion/train2-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_4
    path: AYSLAX/Llama-3.2-1B/dp-diffusion/train4-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_inf
    path: AYSLAX/Llama-3.2-1B/dp-diffusion/trainnodp-dp-diffusion.jsonl
  - split: llama_3_2_1b__epsvec__eps_0_5
    path: AYSLAX/Llama-3.2-1B/EPSVec/train0.5-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_1
    path: AYSLAX/Llama-3.2-1B/EPSVec/train1-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_2
    path: AYSLAX/Llama-3.2-1B/EPSVec/train2-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_4
    path: AYSLAX/Llama-3.2-1B/EPSVec/train4-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_inf
    path: AYSLAX/Llama-3.2-1B/EPSVec/trainnodp-EPSVec.jsonl
  - split: llama_3_2_1b_instruct__dp_transformers__eps_0_5
    path: AYSLAX/Llama-3.2-1B-Instruct/dp-transformers/train0.5-dp-transformers-Llama-3.2-1B-Instruct.jsonl
  - split: llama_3_2_1b_instruct__dp_transformers__eps_1
    path: AYSLAX/Llama-3.2-1B-Instruct/dp-transformers/train1-dp-transformers-Llama-3.2-1B-Instruct.jsonl
  - split: llama_3_2_1b_instruct__dp_transformers__eps_2
    path: AYSLAX/Llama-3.2-1B-Instruct/dp-transformers/train2-dp-transformers-Llama-3.2-1B-Instruct.jsonl
  - split: llama_3_2_1b_instruct__dp_transformers__eps_4
    path: AYSLAX/Llama-3.2-1B-Instruct/dp-transformers/train4-dp-transformers-Llama-3.2-1B-Instruct.jsonl
  - split: llama_3_2_1b_instruct__dp_transformers__eps_inf
    path: AYSLAX/Llama-3.2-1B-Instruct/dp-transformers/trainnodp-dp-transformers-Llama-3.2-1B-Instruct.jsonl
  - split: llama_3_2_3b_instruct__dp_transformers__eps_0_5
    path: AYSLAX/Llama-3.2-3B-Instruct/dp-transformers/train0.5-dp-transformers-Llama-3.2-3B-Instruct.jsonl
  - split: llama_3_2_3b_instruct__dp_transformers__eps_1
    path: AYSLAX/Llama-3.2-3B-Instruct/dp-transformers/train1-dp-transformers-Llama-3.2-3B-Instruct.jsonl
  - split: llama_3_2_3b_instruct__dp_transformers__eps_2
    path: AYSLAX/Llama-3.2-3B-Instruct/dp-transformers/train2-dp-transformers-Llama-3.2-3B-Instruct.jsonl
  - split: llama_3_2_3b_instruct__dp_transformers__eps_4
    path: AYSLAX/Llama-3.2-3B-Instruct/dp-transformers/train4-dp-transformers-Llama-3.2-3B-Instruct.jsonl
  - split: llama_3_2_3b_instruct__dp_transformers__eps_inf
    path: AYSLAX/Llama-3.2-3B-Instruct/dp-transformers/trainnodp-dp-transformers-Llama-3.2-3B-Instruct.jsonl
  - split: llama_3_1_8b_instruct__dp_transformers__eps_0_5
    path: AYSLAX/Llama-3.1-8B-Instruct/dp-transformers/train0.5-dp-transformers-Llama-3.1-8B-Instruct.jsonl
  - split: llama_3_1_8b_instruct__dp_transformers__eps_1
    path: AYSLAX/Llama-3.1-8B-Instruct/dp-transformers/train1-dp-transformers-Llama-3.1-8B-Instruct.jsonl
  - split: llama_3_1_8b_instruct__dp_transformers__eps_2
    path: AYSLAX/Llama-3.1-8B-Instruct/dp-transformers/train2-dp-transformers-Llama-3.1-8B-Instruct.jsonl
  - split: llama_3_1_8b_instruct__dp_transformers__eps_4
    path: AYSLAX/Llama-3.1-8B-Instruct/dp-transformers/train4-dp-transformers-Llama-3.1-8B-Instruct.jsonl
  - split: llama_3_1_8b_instruct__dp_transformers__eps_inf
    path: AYSLAX/Llama-3.1-8B-Instruct/dp-transformers/trainnodp-dp-transformers-Llama-3.1-8B-Instruct.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_inf
    path: AYSLAX/OLMo-2-0425-1B-10epochs/dp-transformers/train0-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_0_5
    path: AYSLAX/OLMo-2-0425-1B-10epochs/dp-transformers/train0.5-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_1
    path: AYSLAX/OLMo-2-0425-1B-10epochs/dp-transformers/train1-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_2
    path: AYSLAX/OLMo-2-0425-1B-10epochs/dp-transformers/train2-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_4
    path: AYSLAX/OLMo-2-0425-1B-10epochs/dp-transformers/train4-dp-transformers.jsonl
- config_name: EURLEX
  data_files:
  - split: train
    path: EURLEX/original/train-original.jsonl
  - split: test
    path: EURLEX/original/test-original.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_0_5
    path: EURLEX/Llama-3.2-1B/dp-transformers/train0.5-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_1
    path: EURLEX/Llama-3.2-1B/dp-transformers/train1-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_2
    path: EURLEX/Llama-3.2-1B/dp-transformers/train2-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_4
    path: EURLEX/Llama-3.2-1B/dp-transformers/train4-dp-transformers.jsonl
  - split: llama_3_2_1b__dp_transformers__eps_inf
    path: EURLEX/Llama-3.2-1B/dp-transformers/trainnodp-dp-transformers.jsonl
  - split: llama_3_2_1b__aug_pe__eps_0_5
    path: EURLEX/Llama-3.2-1B/aug-pe/train0.5-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_1
    path: EURLEX/Llama-3.2-1B/aug-pe/train1-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_2
    path: EURLEX/Llama-3.2-1B/aug-pe/train2-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_4
    path: EURLEX/Llama-3.2-1B/aug-pe/train4-aug-pe.jsonl
  - split: llama_3_2_1b__aug_pe__eps_inf
    path: EURLEX/Llama-3.2-1B/aug-pe/trainnodp-aug-pe.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_0_5
    path: EURLEX/Llama-3.2-1B/dp-diffusion/train0.5-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_1
    path: EURLEX/Llama-3.2-1B/dp-diffusion/train1-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_2
    path: EURLEX/Llama-3.2-1B/dp-diffusion/train2-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_4
    path: EURLEX/Llama-3.2-1B/dp-diffusion/train4-dp-diffusion.jsonl
  - split: llama_3_2_1b__dp_diffusion__eps_inf
    path: EURLEX/Llama-3.2-1B/dp-diffusion/trainnodp-dp-diffusion.jsonl
  - split: llama_3_2_1b__epsvec__eps_0_5
    path: EURLEX/Llama-3.2-1B/EPSVec/train0.5-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_1
    path: EURLEX/Llama-3.2-1B/EPSVec/train1-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_2
    path: EURLEX/Llama-3.2-1B/EPSVec/train2-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_4
    path: EURLEX/Llama-3.2-1B/EPSVec/train4-EPSVec.jsonl
  - split: llama_3_2_1b__epsvec__eps_inf
    path: EURLEX/Llama-3.2-1B/EPSVec/trainnodp-EPSVec.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_inf
    path: EURLEX/OLMo-2-0425-1B-10epochs/dp-transformers/train0-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_0_5
    path: EURLEX/OLMo-2-0425-1B-10epochs/dp-transformers/train0.5-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_1
    path: EURLEX/OLMo-2-0425-1B-10epochs/dp-transformers/train1-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_2
    path: EURLEX/OLMo-2-0425-1B-10epochs/dp-transformers/train2-dp-transformers.jsonl
  - split: olmo_2_0425_1b_10epochs__dp_transformers__eps_4
    path: EURLEX/OLMo-2-0425-1B-10epochs/dp-transformers/train4-dp-transformers.jsonl
---

# SynBench: A Benchmark for Differentially Private Text Generation

Synthetic text generated with differential privacy (DP) from six domain-specific classification
datasets, together with the original splits and the membership-inference audit (MIA) data used in the
NeurIPS 2026 Evaluations & Datasets paper *SynBench: A Benchmark for Differentially Private Text
Generation*.

Three further benchmark datasets (PsyTAR, N2C2'08, MIMIC) are under data use agreements and are **not**
included; see the paper for how to obtain them.

## Datasets and licenses

Each config is one source dataset. The original records **and** every synthetic dataset derived from them
are released under the license of that source dataset — the licenses differ, so check the one for the
config you use. The repository-level `license: other` reflects this.

| config | name in paper | domain | task | train / test | source | license |
|---|---|---|---|---|---|---|
| `hallmarks_of_cancer` | HoC | biomedical | multi-label | 10301 / 1818 | [bigbio/hallmarks_of_cancer](https://huggingface.co/datasets/bigbio/hallmarks_of_cancer) | `gpl-3.0` |
| `DMSAFN` | DMSAFN | finance | single-label (3 classes) | 3876 / 970 | [Daniel-ML/sentiment-analysis-for-financial-news-v2](https://huggingface.co/datasets/Daniel-ML/sentiment-analysis-for-financial-news-v2) | `cc-by-nc-sa-4.0` |
| `Lc37Fnd` | LC37FND | finance | single-label (3 classes) | 5620 / 1424 | [luckycat37/financial-news-dataset](https://huggingface.co/datasets/luckycat37/financial-news-dataset) | `cc-by-nc-sa-4.0` |
| `ZSTFNS` | ZSTFNS | finance | single-label (3 classes) | 9543 / 2388 | [zeroshot/twitter-financial-news-sentiment](https://huggingface.co/datasets/zeroshot/twitter-financial-news-sentiment) | `mit` |
| `AYSLAX` | ASYLAX | legal | single-label (3 classes) | 7999 / 1000 | [clairebarale/AsyLex](https://huggingface.co/datasets/clairebarale/AsyLex) | `cc-by-nc-sa-4.0` |
| `EURLEX` | EURLEX | legal | multi-label (8 EuroVoc concepts) | 4942 / 1245 | [EURLEX57K](https://huggingface.co/datasets/jonathanli/eurlex) | `cc-by-sa-4.0` |

## Loading

```python
from datasets import load_dataset

ds = load_dataset("Ydsun/SynBench", "DMSAFN")
ds["train"]                                   # real training split
ds["test"]                                    # real test split (held out, never used for generation)
ds["llama_3_2_1b__dp_transformers__eps_1"]    # synthetic data, DP-Gen on Llama-3.2-1B, ε = 1
```

Every record is `{"text": str, "label": list[str]}`; single-label datasets use a one-element list.

### Split names

Synthetic splits are named `<generator>__<method>__eps_<ε>` (lower case, non-alphanumerics → `_`):

- **generator**: `llama_3_2_1b` (main results), `llama_3_2_1b_instruct`, `llama_3_2_3b_instruct`,
  `llama_3_1_8b_instruct` (generator-size study; AYSLAX and DMSAFN only), `olmo_2_0425_1b_10epochs`
  (OLMo-2 replication, DP-Gen only, generator fine-tuned for 10 epochs)
- **method**: `dp_transformers` (DP-Gen), `aug_pe` (AUG-PE), `dp_diffusion` (DP-Dfs), `epsvec` (EPSVec)
- **ε**: `0_5`, `1`, `2`, `4`, or `inf` (no DP). In the raw file names no-DP is `nodp` for Llama and `0` for OLMo-2.

## Repository layout

```
<dataset>/
├── original/                      train-original.jsonl, test-original.jsonl
├── Llama-3.2-1B/
│   ├── dp-transformers/           train{0.5,1,2,4,nodp}-dp-transformers.jsonl
│   ├── aug-pe/                    train{…}-aug-pe.jsonl
│   ├── dp-diffusion/              train{…}-dp-diffusion.jsonl
│   └── EPSVec/                    train{…}-EPSVec.jsonl
├── Llama-3.2-{1B,3B}-Instruct/, Llama-3.1-8B-Instruct/   (AYSLAX, DMSAFN)
│   └── dp-transformers/           train{…}-dp-transformers-<model>.jsonl
├── OLMo-2-0425-1B-10epochs/
│   └── dp-transformers/           train{0,0.5,1,2,4}-dp-transformers.jsonl   (0 = no DP)
└── mia/                           (AYSLAX, DMSAFN)
    ├── vulsamples/                all_outliers.jsonl (MIA targets), outlier_indices.npy, config.json
    ├── D_sample/
    │   ├── split_datasets/        D_p_indices.npy, D_a_indices.npy (private / auxiliary split of train)
    │   ├── sampled_private_datasets/   dataset_{0..99}.jsonl + sampling_labels.json (membership ground truth)
    │   └── sampled_4ref_datasets_outliers/  dataset_{0..3}.jsonl + sampling_labels.json (reference sets)
    └── D_synth/{augpe,dp-transformers}/finalresults/
                                   dataset_{i}_e{ε}_samples.jsonl  (synthetic version of private subset i; e0 = no DP)
```

## Synthetic data: record counts

Most generators produce one record per real training record. Counts differ where generation or
post-processing dropped records (OLMo-2 DP-Gen outputs were de-duplicated on `(text, label)`, keeping the
first occurrence; AUG-PE and DP-Dfs drop empty or invalid generations). "–" = not generated.

**hallmarks_of_cancer**

| generator | method | ε=inf | ε=4 | ε=2 | ε=1 | ε=0.5 |
|---|---|---|---|---|---|---|
| Llama-3.2-1B | dp-transformers | 10301 | 10301 | 10301 | 10301 | 10301 |
| Llama-3.2-1B | aug-pe | 9550 | 9451 | 9497 | 9471 | 9410 |
| Llama-3.2-1B | dp-diffusion | 10300 | 10299 | 10273 | 10067 | 9660 |
| Llama-3.2-1B | EPSVec | 10301 | 10301 | 10301 | 10301 | 10301 |
| OLMo-2-0425-1B-10epochs | dp-transformers | 8287 | 10263 | 10262 | 10226 | 10174 |

**DMSAFN**

| generator | method | ε=inf | ε=4 | ε=2 | ε=1 | ε=0.5 |
|---|---|---|---|---|---|---|
| Llama-3.2-1B | dp-transformers | 3876 | 3876 | 3876 | 3876 | 3876 |
| Llama-3.2-1B | aug-pe | 3687 | 3681 | 3677 | 3693 | 3684 |
| Llama-3.2-1B | dp-diffusion | 3876 | 3876 | 3876 | 3638 | 24 |
| Llama-3.2-1B | EPSVec | 3876 | 3876 | 3876 | 3876 | 3876 |
| Llama-3.2-1B-Instruct | dp-transformers | 3872 | 3876 | 3876 | 3876 | 3876 |
| Llama-3.2-3B-Instruct | dp-transformers | 3857 | 3876 | 3876 | 3876 | 3876 |
| Llama-3.1-8B-Instruct | dp-transformers | 3772 | 3876 | 3876 | 3876 | 3876 |
| OLMo-2-0425-1B-10epochs | dp-transformers | 2875 | 3825 | 3839 | 3832 | 3762 |

**Lc37Fnd**

| generator | method | ε=inf | ε=4 | ε=2 | ε=1 | ε=0.5 |
|---|---|---|---|---|---|---|
| Llama-3.2-1B | dp-transformers | 5349 | 5609 | 5615 | 5616 | 5597 |
| Llama-3.2-1B | aug-pe | 5614 | 5614 | 5614 | 5614 | 5614 |
| Llama-3.2-1B | dp-diffusion | 5620 | 5620 | 5620 | 5620 | 5620 |
| Llama-3.2-1B | EPSVec | 5620 | 5620 | 5620 | 5620 | 5620 |
| OLMo-2-0425-1B-10epochs | dp-transformers | 5620 | 5608 | 5617 | 5606 | 5608 |

**ZSTFNS**

| generator | method | ε=inf | ε=4 | ε=2 | ε=1 | ε=0.5 |
|---|---|---|---|---|---|---|
| Llama-3.2-1B | dp-transformers | 6710 | 9435 | 9370 | 9001 | 9541 |
| Llama-3.2-1B | aug-pe | 9010 | 8911 | 8850 | 8836 | 8778 |
| Llama-3.2-1B | dp-diffusion | 9537 | 9543 | 9543 | 9437 | 899 |
| Llama-3.2-1B | EPSVec | 9543 | 9543 | 9543 | 9543 | 9543 |
| OLMo-2-0425-1B-10epochs | dp-transformers | 7353 | 9535 | 9522 | 9525 | 9511 |

**AYSLAX**

| generator | method | ε=inf | ε=4 | ε=2 | ε=1 | ε=0.5 |
|---|---|---|---|---|---|---|
| Llama-3.2-1B | dp-transformers | 8999 | 8994 | 8995 | 8989 | 8998 |
| Llama-3.2-1B | aug-pe | 7958 | 7734 | 7730 | 7713 | 7746 |
| Llama-3.2-1B | dp-diffusion | 7999 | 7999 | 7999 | 7999 | 7999 |
| Llama-3.2-1B | EPSVec | 7999 | 7999 | 7999 | 7999 | 7999 |
| Llama-3.2-1B-Instruct | dp-transformers | 7999 | 7999 | 7999 | 7999 | 7999 |
| Llama-3.2-3B-Instruct | dp-transformers | 7999 | 7999 | 7999 | 7999 | 7999 |
| Llama-3.1-8B-Instruct | dp-transformers | 7999 | 7999 | 7996 | 7999 | 7999 |
| OLMo-2-0425-1B-10epochs | dp-transformers | 7999 | 7997 | 7994 | 7995 | 7995 |

**EURLEX**

| generator | method | ε=inf | ε=4 | ε=2 | ε=1 | ε=0.5 |
|---|---|---|---|---|---|---|
| Llama-3.2-1B | dp-transformers | 4942 | 4942 | 4942 | 4942 | 4942 |
| Llama-3.2-1B | aug-pe | 4858 | 4857 | 4858 | 4856 | 4858 |
| Llama-3.2-1B | dp-diffusion | 4942 | 4561 | 4082 | 2768 | 380 |
| Llama-3.2-1B | EPSVec | 4942 | 4942 | 4942 | 4942 | 4942 |
| OLMo-2-0425-1B-10epochs | dp-transformers | 4933 | 4939 | 4938 | 4938 | 4934 |

## How the data was generated

- **DP-Gen** (`dp-transformers`): the generator is fine-tuned with DP-SGD conditioned on the label as a
  control code, then sampled with one prompt per real training record's label set.
- **AUG-PE** (`aug-pe`): private evolution with the generator as the API model.
- **DP-Dfs** (`dp-diffusion`): DP-SGD-trained text diffusion with a GENIE backbone (independent of the
  Llama/OLMo generator; stored under `Llama-3.2-1B/` for layout only).
- **EPSVec** (`EPSVec`): steering with a privatised dataset vector.

ε ∈ {0.5, 1, 2, 4} plus a no-DP run. Full hyper-parameters are in the paper's appendix A.3.

### Dataset-specific notes

- **EURLEX** is a subset of EURLEX57K restricted to 8 EuroVoc concepts (see the paper for the selection).
- **LC37FND** labels are derived from the sentiment scores of `luckycat37/financial-news-dataset`
  (articles scraped from finance.yahoo.com); see the paper for the mapping.
- **ASYLAX** original texts contain mis-decoded UTF-8 (e.g. `’`); they are released unchanged because all
  generators and classifiers were run on them.
- **DP-Dfs** collapses at small ε on some datasets (e.g. DMSAFN ε=0.5: 24 records; EURLEX ε=0.5: 380;
  ZSTFNS ε=0.5: 899). These files are released as produced; the paper reports such settings as unusable.
- AYSLAX `Llama-3.2-1B/dp-transformers` files have ~8,990–8,999 records (more than the 7,999 training
  records); they are released as produced.
- AYSLAX `mia/D_synth/augpe/finalresults` lacks `dataset_1_e0.5` and `dataset_1_e2`.
- `mia/D_synth/augpe` files originally stored `label` as a string; in this release it is wrapped in a
  one-element list to match every other file. Texts are unchanged.
- OLMo-2 10-epoch DP-Gen files had an extra `prompt` column equal to the generation control code
  (`"\t".join(label) + "\n\n"`); it is dropped here so all splits share the `text`/`label` schema.

## Intended use

Benchmarking DP synthetic text generators (utility: train a classifier on a synthetic split, evaluate on
`test`; fidelity: compare synthetic to real text; privacy: re-run the MIA audit). The synthetic data is
**not** a privacy-safe substitute for the originals here — the originals are public and are released
alongside it.

## Citation

```bibtex
@inproceedings{sun2026synbench,
  title     = {SynBench: A Benchmark for Differentially Private Text Generation},
  author    = {Sun, Yidan and Schlegel, Viktor and Nandakumar, Srinivasan and Zahid, Iqra and Wu, Yuping and
               Wu, Yulong and Li, Hao and Zhang, Jie and Del-Pinto, Warren and Nenadic, Goran and
               Lam, Siew Kei and Bharath, Anil Anthony},
  booktitle = {Advances in Neural Information Processing Systems (Evaluations and Datasets Track)},
  year      = {2026}
}
```

Please also cite the source dataset of every config you use (see the table above).