Skip to content

FEAT: Add sibling loaders for dataset variants pulling distinct upstream artifacts #1789

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
Roman Lutz (romanlutz) wants to merge 5 commits into from
Closed

FEAT: Add sibling loaders for dataset variants pulling distinct upstream artifacts #1789

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
5cb8ae3
Add sibling loaders for dataset variants pulling distinct upstream ar…
Copilot May 23, 2026
c36f446
Rename parent dataset_names to match their pinned default variant
romanlutz May 23, 2026
af6e391
Add DeprecationWarning aliases for renamed dataset names
romanlutz May 23, 2026
01fc7e8
Merge branch 'main' into romanlutz/dataset-config-test-audit
romanlutz May 31, 2026
f097fb1
Merge branch 'main' into romanlutz/dataset-config-test-audit
romanlutz Jun 1, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
37 changes: 31 additions & 6 deletions doc/code/datasets/1_loading_datasets.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,10 +80,20 @@
" 'airt_scams',\n",
" 'airt_sexual',\n",
" 'airt_violence',\n",
" 'aya_redteaming',\n",
" 'babelscape_alert',\n",
" 'aya_redteaming_arabic',\n",
" 'aya_redteaming_english',\n",
" 'aya_redteaming_french',\n",
" 'aya_redteaming_hindi',\n",
" 'aya_redteaming_russian',\n",

@richlundeen Richard Lundeen (richlundeen) Jun 1, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Instead of doing this by name like this, should we make a plan so we can use parameters? It'd be nice to load datasets based on various custom parameters (similar to how we do for initializers and scenarios)

" 'aya_redteaming_serbian',\n",
" 'aya_redteaming_spanish',\n",
" 'aya_redteaming_tagalog',\n",
" 'babelscape_alert_adversarial',\n",
" 'babelscape_alert_original',\n",
" 'beaver_tails',\n",
" 'categorical_harmful_qa',\n",
" 'categorical_harmful_qa_chinese',\n",
" 'categorical_harmful_qa_english',\n",
" 'categorical_harmful_qa_vietnamese',\n",
" 'cbt_bench',\n",
" 'ccp_sensitive_prompts',\n",
" 'coconot_contrast',\n",
Expand All @@ -99,7 +109,8 @@
" 'harmbench',\n",
" 'harmbench_multimodal',\n",
" 'harmful_qa',\n",
" 'hixstest',\n",
" 'hixstest_english',\n",
" 'hixstest_hindi',\n",
" 'jbb_behaviors',\n",
" 'librai_do_not_answer',\n",
" 'llm_lat_harmful',\n",
Expand All @@ -109,6 +120,16 @@
" 'mlcommons_ailuminate',\n",
" 'moral_integrity_corpus',\n",
" 'msts',\n",
" 'msts_arabic',\n",
" 'msts_chinese',\n",
" 'msts_farsi',\n",
" 'msts_french',\n",
" 'msts_german',\n",
" 'msts_hindi',\n",
" 'msts_italian',\n",
" 'msts_korean',\n",
" 'msts_russian',\n",
" 'msts_spanish',\n",
" 'multilingual_vulnerability',\n",
" 'or_bench_80k',\n",
" 'or_bench_hard',\n",
Expand All @@ -118,7 +139,9 @@
" 'psfuzz_steal_system_prompt',\n",
" 'pyrit_example_dataset',\n",
" 'red_team_social_bias',\n",
" 'salad_bench',\n",
" 'salad_bench_attack_enhanced',\n",
" 'salad_bench_base',\n",
" 'salad_bench_defense_enhanced',\n",
" 'sgxstest',\n",
" 'simple_safety_tests',\n",
" 'siuo',\n",
Expand All @@ -129,7 +152,9 @@
" 'toxic_chat',\n",
" 'transphobia_awareness',\n",
" 'visual_leak_bench',\n",
" 'vlguard',\n",
" 'vlguard_safe_safes',\n",
" 'vlguard_safe_unsafes',\n",
" 'vlguard_unsafes',\n",
" 'xstest']"
]
},
Expand Down
54 changes: 54 additions & 0 deletions pyrit/datasets/seed_datasets/remote/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,16 +17,28 @@
_AgentThreatRulesDataset,
) # noqa: F401
from pyrit.datasets.seed_datasets.remote.aya_redteaming_dataset import (
AyaHarmCategory,
AyaLanguage,
_AyaRedteamingArabicDataset,
_AyaRedteamingDataset,
_AyaRedteamingFrenchDataset,
_AyaRedteamingHindiDataset,
_AyaRedteamingRussianDataset,
_AyaRedteamingSerbianDataset,
_AyaRedteamingSpanishDataset,
_AyaRedteamingTagalogDataset,
) # noqa: F401
from pyrit.datasets.seed_datasets.remote.babelscape_alert_dataset import (
_BabelscapeAlertDataset,
_BabelscapeAlertOriginalDataset,
) # noqa: F401
from pyrit.datasets.seed_datasets.remote.beaver_tails_dataset import (
_BeaverTailsDataset,
) # noqa: F401
from pyrit.datasets.seed_datasets.remote.categorical_harmful_qa_dataset import (
_CategoricalHarmfulQAChineseDataset,
_CategoricalHarmfulQADataset,
_CategoricalHarmfulQAVietnameseDataset,
) # noqa: F401
from pyrit.datasets.seed_datasets.remote.cbt_bench_dataset import (
_CBTBenchDataset,
Expand Down Expand Up @@ -74,6 +86,7 @@
from pyrit.datasets.seed_datasets.remote.hixstest_dataset import (
HiXSTestLanguage,
_HiXSTestDataset,
_HiXSTestEnglishDataset,
) # noqa: F401
from pyrit.datasets.seed_datasets.remote.jbb_behaviors_dataset import (
_JBBBehaviorsDataset,
Expand All @@ -94,7 +107,17 @@
_MICDataset,
) # noqa: F401
from pyrit.datasets.seed_datasets.remote.msts_dataset import (
_MSTSArabicDataset,
_MSTSChineseDataset,
_MSTSDataset,
_MSTSFarsiDataset,
_MSTSFrenchDataset,
_MSTSGermanDataset,
_MSTSHindiDataset,
_MSTSItalianDataset,
_MSTSKoreanDataset,
_MSTSRussianDataset,
_MSTSSpanishDataset,
) # noqa: F401
from pyrit.datasets.seed_datasets.remote.multilingual_vulnerability_dataset import ( # noqa: F401
_MultilingualVulnerabilityDataset,
Expand All @@ -119,7 +142,9 @@
_RemoteDatasetLoader,
)
from pyrit.datasets.seed_datasets.remote.salad_bench_dataset import (
_SaladBenchAttackEnhancedDataset,
_SaladBenchDataset,
_SaladBenchDefenseEnhancedDataset,
) # noqa: F401
from pyrit.datasets.seed_datasets.remote.sgxstest_dataset import (
SGXSTestLabel,
Expand Down Expand Up @@ -160,6 +185,8 @@
VLGuardSubcategory,
VLGuardSubset,
_VLGuardDataset,
_VLGuardSafeSafesDataset,
_VLGuardSafeUnsafesDataset,
) # noqa: F401
from pyrit.datasets.seed_datasets.remote.vlsu_multimodal_dataset import (
_VLSUMultimodalDataset,
Expand All @@ -169,6 +196,8 @@
) # noqa: F401

__all__ = [
"AyaHarmCategory",
"AyaLanguage",
"CoCoNotCategory",
"CoCoNotSplit",
"FigStepCategory",
Expand All @@ -186,12 +215,22 @@
"ATRDetectionField",
"ATRVariationType",
"_AgentThreatRulesDataset",
"_AyaRedteamingArabicDataset",
"_AyaRedteamingDataset",
"_AyaRedteamingFrenchDataset",
"_AyaRedteamingHindiDataset",
"_AyaRedteamingRussianDataset",
"_AyaRedteamingSerbianDataset",
"_AyaRedteamingSpanishDataset",
"_AyaRedteamingTagalogDataset",
"_BabelscapeAlertDataset",
"_BabelscapeAlertOriginalDataset",
"_BeaverTailsDataset",
"_CBTBenchDataset",
"_CCPSensitivePromptsDataset",
"_CategoricalHarmfulQAChineseDataset",
"_CategoricalHarmfulQADataset",
"_CategoricalHarmfulQAVietnameseDataset",
"_CoCoNotContrastDataset",
"_CoCoNotRefusalDataset",
"_ComicJailbreakDataset",
Expand All @@ -206,13 +245,24 @@
"_HarmBenchMultimodalDataset",
"_HarmfulQADataset",
"_HiXSTestDataset",
"_HiXSTestEnglishDataset",
"_JBBBehaviorsDataset",
"_LibrAIDoNotAnswerDataset",
"_LLMLatentAdversarialTrainingDataset",
"_MedSafetyBenchDataset",
"_MICDataset",
"_MLCommonsAILuminateDataset",
"_MSTSArabicDataset",
"_MSTSChineseDataset",
"_MSTSDataset",
"_MSTSFarsiDataset",
"_MSTSFrenchDataset",
"_MSTSGermanDataset",
"_MSTSHindiDataset",
"_MSTSItalianDataset",
"_MSTSKoreanDataset",
"_MSTSRussianDataset",
"_MSTSSpanishDataset",
"_MultilingualVulnerabilityDataset",
"_ORBench80KDataset",
"_ORBenchHardDataset",
Expand All @@ -222,7 +272,9 @@
"_RedTeamSocialBiasDataset",
"_RemoteDatasetLoader",
"_SGXSTestDataset",
"_SaladBenchAttackEnhancedDataset",
"_SaladBenchDataset",
"_SaladBenchDefenseEnhancedDataset",
"_SimpleSafetyTestsDataset",
"_SIUODataset",
"_SOSBenchDataset",
Expand All @@ -232,6 +284,8 @@
"_ToxicChatDataset",
"_TransphobiaAwarenessDataset",
"_VLGuardDataset",
"_VLGuardSafeSafesDataset",
"_VLGuardSafeUnsafesDataset",
"_VLSUMultimodalDataset",
"_VisualLeakBenchDataset",
"VisualLeakBenchCategory",
Expand Down
Loading
Loading