Skip to content

FIX: match XL-SafetyBench cultural category labels to upstream #3006

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

Open
Edson (Zhuoxi2000) wants to merge 2 commits into
base: main
Choose a base branch
from
+39 −3
Open

FIX: match XL-SafetyBench cultural category labels to upstream #3006

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d78d914
FIX: match XL-SafetyBench cultural category labels to upstream
Zhuoxi2000 Oct 1, 2026
4e0951b
Merge branch 'main' into fix-xlsb-cultural-category-commas
hannahwestra25 Oct 9, 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
6 changes: 3 additions & 3 deletions pyrit/datasets/seed_datasets/remote/xl_safety_bench_dataset.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -89,10 +89,10 @@ class XLSafetyBenchCulturalCategory(str, Enum):
"""Cultural categories used by the XL-SafetyBench Cultural track."""

SYMBOLIC_TABOOS_AND_GIFT_GIVING = "Symbolic Taboos & Gift-Giving"
FOOD_DIETARY_LAW_AND_HOSPITALITY = "Food Dietary Law & Hospitality"
DEATH_GRIEF_AND_FUNERAL_PRACTICES = "Death Grief & Funeral Practices"
FOOD_DIETARY_LAW_AND_HOSPITALITY = "Food, Dietary Law & Hospitality"
DEATH_GRIEF_AND_FUNERAL_PRACTICES = "Death, Grief & Funeral Practices"
DAILY_LIFE_AND_PUBLIC_CONDUCT = "Daily Life & Public Conduct"
HIERARCHY_ADDRESS_AND_SOCIAL_DEFERENCE = "Hierarchy Address & Social Deference"
HIERARCHY_ADDRESS_AND_SOCIAL_DEFERENCE = "Hierarchy, Address & Social Deference"
LEGAL_LANDMINES = "Legal Landmines"


Expand Down
36 changes: 36 additions & 0 deletions tests/unit/datasets/test_xl_safety_bench_dataset.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -294,6 +294,42 @@ async def test_cultural_category_filter_and_hidden_violation():
assert seed.metadata["scenario_english"] == "English scenario 2 for japan"


@pytest.mark.parametrize(
"category, upstream_label",
[
# Labels exactly as they appear in the ``category`` column of every
# data/cultural/<country>/scenario_prompts.csv on the Hugging Face dataset.
(XLSafetyBenchCulturalCategory.FOOD_DIETARY_LAW_AND_HOSPITALITY, "Food, Dietary Law & Hospitality"),
(XLSafetyBenchCulturalCategory.DEATH_GRIEF_AND_FUNERAL_PRACTICES, "Death, Grief & Funeral Practices"),
(
XLSafetyBenchCulturalCategory.HIERARCHY_ADDRESS_AND_SOCIAL_DEFERENCE,
"Hierarchy, Address & Social Deference",
),
],
)
async def test_cultural_category_filter_matches_upstream_labels_with_commas(category, upstream_label):
loader = _XLSafetyBenchCulturalDataset(countries=[XLSafetyBenchCountry.FRANCE], categories=[category])
rows = [
{
"id": "CC-fr-1",
"category": upstream_label,
"sensitivity_english": "s",
"sensitivity_local": "s",
"base_query_english": "q",
"base_query_local": "q",
"scenario_english": "English scenario",
"scenario_local": "Local scenario",
"hidden_violation": "v",
}
]

with patch.object(loader, "_fetch_from_url", return_value=rows):
dataset = await loader.fetch_dataset_async()

assert len(dataset.seeds) == 1
assert dataset.seeds[0].metadata["category"] == upstream_label


async def test_cultural_skips_empty_scenario():
loader = _XLSafetyBenchCulturalDataset(countries=[XLSafetyBenchCountry.FRANCE])
rows = [
Expand Down