← Evaluation overview 9 research contributions
Safety & Factuality Factual knowledge, multimodal hallucination, jailbreak resistance, over-refusal, and human-centered behavior.
Benchmarks & related work 9 works
Search by name or capability, or narrow by publication status and authorship.
Clear filters
Factuality & hallucination
OmniHalluc-L
Evaluates long-form omni-modal hallucination using counterfactual and modality-perturbation analysis.
Hallucination Audio-visual
EMNLP 2026 · Accepted Corresponding author
Paper & authors OmniHalluc-L: Counterfactual Benchmarking and Modality-Perturbation Calibration for Long-Form Omni Hallucination
Zixuan Dong, Jiafu Tang, Zhide Lei, Zhe Cao, Zijie Zhang, Yanghai Wang, Shihao Li, Xiaodong Wang, Baoyun Peng, Jiaheng Liu#
* Equal contribution; # Corresponding author.
Factuality & hallucination
MSQA
Evaluates multilingual and multicultural factual knowledge with natively sourced questions.
Factuality Multilingual
Findings of EMNLP 2026 · Accepted
Paper & authors MSQA: A Natively Sourced Multilingual and Multicultural SimpleQA Benchmark
Yukai Huang, Xianru Chen, Mingxiang Chen, Xinping Lei, Fangbing Deng, Jin Chen, Jiaheng Liu , Ge Zhang, Wenhao Huang
* Equal contribution; # Corresponding author.
Safety & human alignment
USB
Provides unified evaluation of multimodal safety and over-refusal.
Multimodal safety Over-refusal
ACL 2026
Paper & authors USB: A Comprehensive and Unified Safety Evaluation Benchmark for Multimodal Large Language Models
Baolin Zheng# , Guanlin Chen, Qingyang Teng, Hongqiong Zhong, Yingshui Tan, Zhendong Liu, Weixun Wang, Jiaheng Liu , Jian Yang, Huiyun Jing, Jincheng Wei, Wenbo Su, Xiaoyong Zhu, Bo Zheng, Kaifu Zhang
* Equal contribution; # Corresponding author.
Safety & human alignment
SafeDialBench
Evaluates safety in multi-turn dialogues with diverse jailbreak attacks.
Jailbreak resistance Multi-turn
ICLR 2026
Paper & authors SafeDialBench: A Fine-Grained Safety Evaluation Benchmark for Large Language Models in Multi-Turn Dialogues with Diverse Jailbreak Attacks
Hongye Cao, Sijia Jing, Yanming Wang, Ziyue Peng, Zhixin Bai, Zhe Cao, Meng Fang, Fan Feng, Jiaheng Liu , Boyan Wang, Tianpei Yang, Jing Huo# , Yang Gao, Fanyu Meng# , Xi Yang, Chao Deng, Junlan Feng
* Equal contribution; # Corresponding author.
Safety & human alignment Companion dataset
FigStep-benign
Evaluates over-refusal on benign multimodal requests constructed in the DREAM study.
Over-refusal Multimodal safety
NAACL 2025
Paper & authors DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models
Jianyu Liu, Hangyu Guo, Ranjie Duan, Xingyuan Bu# , Yancheng He, Shilong Li, Hui Huang, Jiaheng Liu , Yucheng Wang, Chenchen Jing, Xingwei Qu, Xiao Zhang, Pei Wang, Yanan Wu, Jihao Gu, Yangguang Li, Jianke Zhu
* Equal contribution; # Corresponding author.
Safety & human alignment
Chinese SafetyQA
Tests the factual accuracy of safety-related knowledge.
Safety knowledge Factuality
ACL 2025
Paper & authors Chinese SafetyQA: A Safety Short-form Factuality Benchmark for Large Language Models
Yingshui Tan# , Boren Zheng, Baihui Zheng, Kerui Cao, Huiyun Jing, Jincheng Wei, Jiaheng Liu , Yancheng He, Wenbo Su, Xiaoyong Zhu, Bo Zheng, Kaifu Zhang
* Equal contribution; # Corresponding author.
Factuality & hallucination
Chinese SimpleQA
Evaluates short-form factual question answering in Chinese.
Factuality Chinese
ACL 2025 Corresponding author
Paper & authors Chinese SimpleQA: A Chinese Factuality Evaluation for Large Language Models
Yancheng He, Shilong Li, Jiaheng Liu# , Yingshui Tan, Weixun Wang, Hui Huang, Xingyuan Bu, Hangyu Guo, Chengwei Hu, Boren Zheng, Zhuoran Lin, Dekai Sun, Zhicheng Zheng, Wenbo Su, Bo Zheng
* Equal contribution; # Corresponding author.
Safety & human alignment
BSA Bench（Beyond Safe Answers）
Tests whether reasoning models recognize risk beyond producing superficially safe answers.
Risk awareness Safety reasoning
arXiv preprint, 2025
Paper & authors Beyond Safe Answers: A Benchmark for Evaluating True Risk Awareness in Large Reasoning Models
Baihui Zheng, Boren Zheng, Kerui Cao, Yingshui Tan# , Zhendong Liu, Weixun Wang, Jiaheng Liu , Jian Yang, Wenbo Su, Xiaoyong Zhu, Bo Zheng, Kaifu Zhang
* Equal contribution; # Corresponding author.
Safety & human alignment
GIEBench
Evaluates empathy through the lens of group identity.
Empathy Human alignment
arXiv preprint, 2024
Paper & authors GIEBench: Towards Holistic Evaluation of Group Identity-based Empathy for Large Language Models
Leyan Wang, Yonggang Jin, Tianhao Shen, Tianyu Zheng, Xinrun Du, Chenchen Zhang, Wenhao Huang, Jiaheng Liu , Shi Wang, Ge Zhang# , Liuyu Xiang# , Zhaofeng He
* Equal contribution; # Corresponding author.
No matching work Try a broader keyword or clear the filters.
Show more
This collection includes co-authored benchmarks, companion evaluation datasets, and evaluation methods. Each work has one primary category; topic tags capture related capabilities. Corresponding-author labels refer to Jiaheng Liu.
Jiaheng Liu · Nanjing University Updated September 2026
Jiaheng LiuNANJING UNIVERSITY · LINK LAB