Skip to content

ifeval_ood: read the public IFBench_test dataset #326

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

Merged
abhishekraok merged 2 commits into from
Aug 27, 2026
Merged

ifeval_ood: read the public IFBench_test dataset #326

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
204a40c
ifeval_ood: read the public IFBench_test dataset
abhishekraok Aug 25, 2026
83da0aa
Merge branch 'main' into fix/ifbench-public-dataset
abhishekraok Aug 27, 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
2 changes: 1 addition & 1 deletion src/olmo_eval/common/scorers/ifeval.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
"""Scorer for IFBench / IFEval instruction-following evaluation.

Uses the ``ifbench`` package registry, which covers the original IFEval
(DEFAULT) verifiers, the OOD verifiers used by ``allenai/IFBench_test2``,
(DEFAULT) verifiers, the OOD verifiers used by ``allenai/IFBench_test``,
and the verifiers used by the multi-turn ``VGraf/ifeval_mt`` slices. The
scorer evaluates a response against per-instance instructions (looked up
in ``instance.metadata["instruction_id_list"]`` / ``"kwargs"``) and writes
Expand Down
4 changes: 2 additions & 2 deletions src/olmo_eval/evals/tasks/ifeval_ood.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
"""IFEval OOD: out-of-distribution instruction-following slice of IFBench.

Dataset: ``allenai/IFBench_test2`` (300 prompts), each carrying a list of
Dataset: ``allenai/IFBench_test`` (300 prompts), each carrying a list of
instruction IDs and per-instruction kwargs. Verifiers come from the vendored
registry in :mod:`olmo_eval.common.scorers.ifeval_deps`.

Expand Down Expand Up @@ -35,7 +35,7 @@

@register("ifeval_ood")
class IFEvalOOD(Task):
data_source = DataSource(path="allenai/IFBench_test2", split="train")
data_source = DataSource(path="allenai/IFBench_test", split="train")
split = Split.TRAIN
metrics = (
IFEvalPromptStrictAccuracy(),
Expand Down
Loading