Saved in:
Bibliographic Details
Main Authors: Hasan, Alif Al, Biswas, Sumon
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.05427
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911735611392000
author Hasan, Alif Al
Biswas, Sumon
author_facet Hasan, Alif Al
Biswas, Sumon
contents Refusal rates are a poor proxy for LLM safety, i.e., a model may over-refuse benign prompts while still complying with harmful ones. We audit both failure modes across 21 open-weight LLMs on four safety benchmarks (OR-Bench, XSTest, ToxiGen, BOLD), using a composition adjustment to isolate model sensitivity from dataset toxicity confounds. We report three findings. First, models adopt fundamentally different calibration strategies: conservative ecosystems such as Llama suppress unsafe outputs at the cost of elevated over-refusals, while permissive ecosystems such as DeepSeek and Qwen preserve helpfulness but tolerate higher harmful compliance. Second, demographic protection is unequal: models over-protect prominent racial and religious groups, frequently refusing even benign prompts about them, while providing substantially weaker protection against disability-targeted attacks. Third, refusal and compliance tendencies are stable within model families across generations and scales, suggesting that post-training objectives shape safety behavior more than architecture. Our results call for joint, demographically-aware, and multi-judge safety evaluation.
format Preprint
id arxiv_https___arxiv_org_abs_2605_05427
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle The Refusal--Compliance Tradeoff: A Large-Scale Safety Behavior Audit of Large Language Models
Hasan, Alif Al
Biswas, Sumon
Artificial Intelligence
Refusal rates are a poor proxy for LLM safety, i.e., a model may over-refuse benign prompts while still complying with harmful ones. We audit both failure modes across 21 open-weight LLMs on four safety benchmarks (OR-Bench, XSTest, ToxiGen, BOLD), using a composition adjustment to isolate model sensitivity from dataset toxicity confounds. We report three findings. First, models adopt fundamentally different calibration strategies: conservative ecosystems such as Llama suppress unsafe outputs at the cost of elevated over-refusals, while permissive ecosystems such as DeepSeek and Qwen preserve helpfulness but tolerate higher harmful compliance. Second, demographic protection is unequal: models over-protect prominent racial and religious groups, frequently refusing even benign prompts about them, while providing substantially weaker protection against disability-targeted attacks. Third, refusal and compliance tendencies are stable within model families across generations and scales, suggesting that post-training objectives shape safety behavior more than architecture. Our results call for joint, demographically-aware, and multi-judge safety evaluation.
title The Refusal--Compliance Tradeoff: A Large-Scale Safety Behavior Audit of Large Language Models
topic Artificial Intelligence
url https://arxiv.org/abs/2605.05427