OR-Bench: An Over-Refusal Benchmark for Large Language Models

ICLR 2025 Conference Submission541 Authors

Keywords: over-refusal, llm, safety, false-refusal

TL;DR: an over-refusal benchmark for large language model

Abstract:

Large Language Models (LLMs) require careful safety alignment to prevent malicious outputs. While significant research focuses on mitigating harmful content generation,
the enhanced safety often come with the side effect of over-refusal, where LLMs may reject innocuous prompts and become less helpful.
Although the issue of over-refusal has been empirically observed, a systematic measurement is challenging
due to the difficulty of crafting prompts that can elicit the over-refusal behaviors of LLMs.
This study proposes a novel method for automatically generating large-scale over-refusal datasets. Leveraging this technique, we introduce OR-Bench, the first large-scale over-refusal benchmark. OR-Bench comprises 80,000 over-refusal prompts across 10 common rejection categories, a subset of around 1,000 hard prompts that are challenging even for state-of-the-art LLMs, and an additional 600 toxic prompts to prevent indiscriminate responses.
We then conduct a comprehensive study to measure the over-refusal of 32 popular LLMs across 8 model families.
To facilitate reproducibility, we host our datasets, along with an interactive demo and leaderboard, on HuggingFace at https://huggingface.co/spaces/orbench-llm/or-bench and release our code at https://github.com/orbench/or-bench. We hope this benchmark can help the community develop better safety aligned models.

Primary Area: datasets and benchmarks

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 541