Beyond the Safety Bundle: Auditing the Helpful and Harmless Dataset (2025.naacl-long)
Copied to clipboard
| Challenge: | Learning from human feedback (LHF) has been used to mitigate the harms of large language models (LLMs) but the quality of this feedback and its effectiveness as a safety mitigation technique remain unclear. |
| Approach: | They audit the Helpful and Harmless (HH) dataset by Anthropic and examine how conceptualization failures and quality issues identified in the dataset can create additional harms . |
| Outcome: | The findings highlight the need for more nuanced, context-sensitive approaches to safety mitigation in large language models. |
Similar Papers
From Representational Harms to Quality-of-Service Harms: A Case Study on Llama 2 Safety Safeguards (2024.findings-acl)
Copied to clipboard
Khaoula Chehbouni, Megha Roshan, Emmanuel Ma, Futian Wei, Afaf Taik, Jackie Cheung, Golnoosh Farnadi
| Challenge: | Recent advances in large language models have also introduced additional safety risks and raised concerns regarding their detrimental impact on already marginalized populations. |
| Approach: | They propose to use LLMs to evaluate their safety responses on already mitigated biases by evaluating models on already encoded assumptions. |
| Outcome: | The proposed model can encode harmful assumptions, but it can also be harmful for certain demographic groups. |
PKU-SafeRLHF: Towards Multi-Level Safety Alignment for LLMs with Human Preference (2025.acl-long)
Copied to clipboard
Jiaming Ji, Donghai Hong, Borong Zhang, Boyuan Chen, Josef Dai, Boren Zheng, Tianyi Alex Qiu, Jiayi Zhou, Kaile Wang, Boxun Li, Sirui Han, Yike Guo, Yaodong Yang
| Challenge: | Using large-scale annotation data, large language models can generate noise, errors and biases, leading to unexpected behaviours. |
| Approach: | They propose a dataset to promote safety alignment in large language models . they separate helpfulness and harmlessness annotations for question-answering pairs . |
| Outcome: | The proposed dataset provides 44.6k prompts and 265k question-answer pairs with safety meta-labels for 19 harm categories and three severity levels, with answers generated by Llama-family models. |
Soteria: Language-Specific Functional Parameter Steering for Multilingual Safety Alignment (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Soteria locates and minimally adjusts the “functional heads” most responsible for harmful content generation in each language. |
| Approach: | Soteria locates and minimally adjusts the "functional heads" responsible for harmful content generation in each language. |
| Outcome: | The proposed approach reduces harmful content generation in languages while preserving model performance. |
Safety of Large Language Models Beyond English: A Systematic Literature Review of Risks, Biases, and Safeguards (2026.eacl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) have a growing number of applications that generate harmful, biased, or unsafe content. |
| Approach: | They synthesize findings from recent studies that evaluate their robustness across languages . they highlight gaps in multilingual safety research and recommend future work . |
| Outcome: | The systematic review examines the multilingual safety of large language models in English . it identifies challenges such as dataset availability and evaluation biases . |
XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Models (2024.naacl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) are now being used by millions of people across the world. |
| Approach: | They propose a test suite called XSTest to identify such eXaggerated Safety behaviours in a systematic way. |
| Outcome: | The proposed test suite identifies eXaggerated Safety behaviours in a systematic way. |
Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing surveys focus on interpretation or safety, but safety and understanding are core motivations for interpretation research. |
| Approach: | They propose a framework that connects interpretation methods, enhancements they inform, and tools that operationalize them. |
| Outcome: | The proposed framework summarizes nearly 70 studies at their intersections and concludes with open challenges and future directions. |
The Language Barrier: Dissecting Safety Challenges of LLMs in Multilingual Contexts (2024.findings-acl)
Copied to clipboard
Lingfeng Shen, Weiting Tan, Sihao Chen, Yunmo Chen, Jingyu Zhang, Haoran Xu, Boyuan Zheng, Philipp Koehn, Daniel Khashabi
| Challenge: | Recent studies show that malicious prompt instructions could solicit objectionable content from LLMs. |
| Approach: | They compare how state-of-the-art LLMs respond to malicious prompts in different languages . they find that LLM's generate unsafe responses more often when a prompt is written in a lower-resource language . |
| Outcome: | The proposed model can generate unsafe responses more often when a malicious prompt is written in a lower-resource language, and less irrelevant responses when written in lower-source languages. |
A Chinese Dataset for Evaluating the Safeguards in Large Language Models (2024.findings-acl)
Copied to clipboard
Yuxia Wang, Zenan Zhai, Haonan Li, Xudong Han, Shom Lin, Zhenxuan Zhang, Angela Zhao, Preslav Nakov, Timothy Baldwin
| Challenge: | a recent study has shown that large language models can produce harmful responses, exposing users to unexpected risks. |
| Approach: | They propose a dataset for the safety evaluation of Chinese LLMs in Mandarin Chinese . they extend the dataset to better identify false negative and false positive examples . |
| Outcome: | The proposed dataset is for the safety evaluation of Chinese LLMs, and is based on a Chinese dataset. |
Beyond Reactive Safety: Risk-Aware LLM Alignment via Long-Horizon Simulation (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing alignment methods focus on reactive feedback, where immediate human perception is leveraged to judge sampled model responses as preference data for post-training. |
| Approach: | They propose a proof-of-concept framework that projects how model-generated advice could propagate through societal systems on a macroscopic scale over time, enabling more robust alignment. |
| Outcome: | The proposed framework achieves 20% improvement on existing safety benchmarks and an average win rate exceeding 70% against strong baselines. |
Multitask-Bench: Unveiling and Mitigating Safety Gaps in LLMs Fine-tuning (2025.coling-main)
Copied to clipboard
| Challenge: | Recent advances in Large Language Models (LLMs) have led to their adoption across a wide range of tasks, ranging from code generation to machine translation and sentiment analysis. |
| Approach: | They propose to fine-tune LLMs on benign (non-harmful) data to ensure safe outputs. |
| Outcome: | The proposed model reduces attack success rates across a range of tasks without compromising its usefulness. |