Papers with refusal rates
SAGE: A Generic Framework for LLM Safety Evaluation (2025.emnlp-industry)
Copied to clipboard
| Challenge: | Current safety evaluation methodologies focus on single-turn interactions with generic policies, failing to capture conversational dynamics of real-world usage and application-specific harms. |
| Approach: | They propose a framework for customized and dynamic harm evaluations that employs prompted adversarial agents with diverse personalities based on the Big Five model. |
| Outcome: | The proposed framework enables system-aware multi-turn conversations that adapt to target applications and harm policies. |
Alignment Quality Index (AQI) : Beyond Refusals: AQI as an Intrinsic Alignment Diagnostic via Latent Geometry, Cluster Divergence, and Layer wise Pooled Representations (2025.emnlp-main)
Copied to clipboard
Abhilekh Borah, Chhavi Sharma, Danush Khanna, Utkarsh Bhatt, Gurpreet Singh, Hasnat Md Abdullah, Raghav Kaushik Ravi, Vinija Jain, Jyoti Patel, Shubham Singh, Vasu Sharma, Arpita Vats, Rahul Raja, Aman Chadha, Amitava Das
| Challenge: | a new metric measures the quality of large language models (LLMs) that detects hidden misalignments and jailbreak risks. |
| Approach: | They propose a decoding-invariant metric that measures latent safety failures . they propose 'Alignment Quality Index' to measure latent activations in latent space . |
| Outcome: | The proposed metric detects latent safety failures overlooked by behavioral benchmarks and jailbreaks. |
Characterizing Selective Refusal Bias in Large Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | a recent study shows that safety guardrails in large language models can inadvertently introduce or reflect new biases as they may refuse to generate harmful content targeting some demographic groups and not others. |
| Approach: | They examine the selective refusal bias in large language models by examining demographics and responses. |
| Outcome: | The proposed model fails to defend against an indirect attack on previously refused groups in 89% of the trials. |
Path Drift in Large Reasoning Models: How First-Person Commitments Override Safety (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies on prompt injection and jailbreak attacks primarily target the surface structure of input prompts. |
| Approach: | They propose a three-stage approach to mitigate the risk of Long-CoT reasoning drift . they propose 'path-level defense' strategy that incorporates role attribution correction and metacognitive reflection . |
| Outcome: | The proposed framework reduces refusal rates and ethical evaporation, while ethical escalation and layered disclaimers progressively steer models toward unsafe completions. |
How Jailbreak Defenses Work and Ensemble? A Mechanistic Investigation (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Jailbreak attacks, where harmful prompts bypass generative models’ built-in safety, raise serious concerns about model vulnerability. |
| Approach: | They propose to reframe the standard generation task as a binary classification problem to assess model refusal tendencies for both harmful and benign queries. |
| Outcome: | The proposed defenses improve model safety or optimize the trade-off between safety and helpfulness. |
PsychEthicsBench: Evaluating Large Language Models Against Australian Mental Health Ethics (2026.findings-acl)
Copied to clipboard
Yaling Shen, Stephanie Fong, Yiwen Jiang, Zimu Wang, Feilong Tang, Qingyang Xu, Xiangyu Zhao, Zhongxing Xu, Jiahe Liu, Jinpeng Hu, Dominic Dwyer, Zongyuan Ge
| Challenge: | Mental disorders affect nearly one in seven people worldwide, yet the vast majority do not receive adequate care. |
| Approach: | They propose a framework to evaluate LLMs' ethical knowledge and behavioral responses through multiple-choice and open-ended tasks with fine-grained ethicality annotations. |
| Outcome: | Empirical results across 14 models reveal that refusal rates are poor indicators of ethical behavior, revealing a significant divergence between safety triggers and clinical appropriateness. |