# Beyond Over-Refusal: Scenario-Based Diagnostics and Post-Hoc Mitigation for Exaggerated Refusals in LLMs

arXiv: https://arxiv.org/abs/2510.08158 · HTML: https://mlsanity.com/paper/2510.08158/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 26 of its 40 references that are in the map.
- First posted on arXiv 9 Oct 2025.

## Placed in

- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 26 of 26)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [A Unified Approach to Interpreting Model Predictions](https://mlsanity.com/paper/1705.07874/) (2017)
- ["Why Should I Trust You?": Explaining the Predictions of Any Classifier](https://mlsanity.com/paper/1602.04938/) (2016)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity](https://mlsanity.com/paper/2101.03961/) (2021)
- [Jailbroken: How Does LLM Safety Training Fail?](https://mlsanity.com/paper/2307.02483/) (2023)
- [GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints](https://mlsanity.com/paper/2305.13245/) (2023)
- [RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models](https://mlsanity.com/paper/2009.11462/) (2020)
- [A General Language Assistant as a Laboratory for Alignment](https://mlsanity.com/paper/2112.00861/) (2021)
- [Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotations](https://mlsanity.com/paper/2312.08935/) (2023)
- [Captum: A unified and generic model interpretability library for PyTorch](https://mlsanity.com/paper/2009.07896/) (2020)
- [ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection](https://mlsanity.com/paper/2203.09509/) (2022)
- [DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence](https://mlsanity.com/paper/2406.11931/) (2024)
- [Understanding Neural Networks through Representation Erasure](https://mlsanity.com/paper/1612.08220/) (2016)
- [Imbalanced-learn: A Python Toolbox to Tackle the Curse of Imbalanced Datasets in Machine Learning](https://mlsanity.com/paper/1609.06570/) (2016)
- [Alignment of Language Agents](https://mlsanity.com/paper/2103.14659/) (2021)
- [The Hidden Risks of Large Reasoning Models: A Safety Assessment of R1](https://mlsanity.com/paper/2502.12659/) (2025)
- [Large Language Model Safety: A Holistic Survey](https://mlsanity.com/paper/2412.17686/) (2024)
- [From Hard Refusals to Safe-Completions: Toward Output-Centric Safety Training](https://mlsanity.com/paper/2508.09224/) (2025)
- [SCANS: Mitigating the Exaggerated Safety for LLMs via Safety-Conscious Activation Steering](https://mlsanity.com/paper/2408.11491/) (2024)
- [Evaluating Generative Models for Graph-to-Text Generation](https://mlsanity.com/paper/2307.14712/) (2023)
- [Think Before Refusal : Triggering Safety Reflection in LLMs to Mitigate False Refusal Behavior](https://mlsanity.com/paper/2503.17882/) (2025)
- [CoDAE: Adapting Large Language Models for Education via Chain-of-Thought Data Augmentation](https://mlsanity.com/paper/2508.08386/) (2025)
- [LLM in the Loop: Creating the ParaDeHate Dataset for Hate Speech Detoxification](https://mlsanity.com/paper/2506.01484/) (2025)

## Built on by (most cited first, 3 of 3)

- [AgentAbstain: Do LLM Agents Know When Not to Act?](https://mlsanity.com/paper/2607.10059/) (2026)
- [Analyzing Bias in False Refusal Behavior of Large Language Models for Hate Speech Detoxification](https://mlsanity.com/paper/2601.08668/) (2026)
- [Cyber-Capable AI Agents: Vulnerabilities, Evaluation Containment, and Defensive Response](https://mlsanity.com/paper/2607.25379/) (2026)

## Abstract (arXiv)

Large language models (LLMs) frequently produce false refusals, declining benign requests that contain terms resembling unsafe queries. We address this challenge by introducing two comprehensive benchmarks: the Exaggerated Safety Benchmark (XSB) for single-turn prompts, annotated with "Focus" keywords that identify refusal-inducing triggers, and the Multi-turn Scenario-based Exaggerated Safety Benchmark (MS-XSB), which systematically evaluates refusal calibration in realistic, context-rich dialog settings. Our benchmarks reveal that exaggerated refusals persist across diverse recent LLMs and are especially pronounced in complex, multi-turn scenarios. To mitigate these failures, we leverage post-hoc explanation methods to identify refusal triggers and deploy three lightweight, model-agnostic approaches, ignore-word instructions, prompt rephrasing, and attention steering, at inference time, all without retraining or parameter access. Experiments on four instruction-tuned Llama models demonstrate that these strategies substantially improve compliance on safe prompts while maintaining robust safety protections. Our findings establish a reproducible framework for diagnosing and mitigating exaggerated refusals, highlighting practical pathways to safer and more helpful LLM deployments.