Papers by Yunpeng Liu
InsLogicBench: An Argumentation Logic Grounded Benchmark for Complex Insurance Claims Adjudication (2026.acl-long)
Copied to clipboard
Jin Liu, Yunpeng Liu, Keyi Wang, Jie Shi, Xiao Xu, Wenkang Huang, Xingzhong Xu, Xin Liang, Yanghua Xiao
| Challenge: | Existing benchmarks for insurance claims adjudication are limited to information retrieval or simple multiple-choice setups. |
| Approach: | They propose a benchmark that provides complete reasoning traces linking factual inputs, relevant policy clauses, and final verdicts. |
| Outcome: | The proposed benchmark shows that models often produce correct decisions but fail to provide precise justifications, highlighting a critical discrepancy between decision accuracy and logical reasoning capabilities. |
LitVISTA: A Benchmark for Narrative Orchestration in Literary Text (2026.acl-long)
Copied to clipboard
Mingzhe Lu, Yiwen Wang, Yanbing Liu, Qi You, Chong Liu, Ruize Qin, Haoyu Dong, Wenyu Zhang, JiaRui Zhang, Yue Hu, Yunpeng Li
| Challenge: | Existing large language models focus on causal coherence, neglecting the complex story arcs and orchestration inherent in human narratives. |
| Approach: | They propose a high-dimensional framework for narrative orchestration that unifies human and model perspectives while jointly characterizing narrative function and structure in a common space. |
| Outcome: | The proposed framework unifies human and model perspectives while jointly characterizing narrative function and structure in a common space. |
d-TreeRPO: Towards More Reliable Policy Optimization for Diffusion Language Models (2026.acl-long)
Copied to clipboard
Leyi Pan, Shuchang Tao, Yunpeng Zhai, Zheyu Fu, Liancheng Fang, Minghua He, Lingzhe Zhang, Zhaoyang Liu, Bolin Ding, Aiwei Liu, Lijie Wen
| Challenge: | Existing RL methods suffer from reliability bottlenecks due to reward sparsity and intractable computations . d-TreeRPO provides fine-grained and verifiable step-wise reward signals . |
| Approach: | They propose a reliable reinforcement learning framework for diffusion large language models that leverages tree-structured rollouts and bottom-up advantage computation based on verifiable outcome rewards. |
| Outcome: | The proposed framework outperforms baseline models and achieves significant improvements across reasoning benchmarks. |
Don’t Corrupt the Fact: A Trustworthy RAG Watermarking Framework based on Dual Factual Shield (2026.acl-long)
Copied to clipboard
| Challenge: | Existing watermarking methods are fact-agnostic and cause "faithfulness hallucinations" a novel framework to enforce knowledge loyalty is proposed to improve watermarks . |
| Approach: | They propose a new framework that enforces knowledge loyalty by spoofing terms from retrieved contexts and prompt-based semantic guidance to protect against factual corruption. |
| Outcome: | The proposed framework reduces the Knowledge Corruption Rate while maintaining its original high security and robustness. |
Can We Steer Reasoning Direction by Thinking Intervention? (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Large Reason Models suffer from overthinking and erroneous reasoning problems due to the lack of fine-grained control over their reasoning behaviors. |
| Approach: | They propose a paradigm to enable fine-grained control over LRMs’ reasoning behaviors by aligning reasoning trajectories with specific cognitive patterns. |
| Outcome: | The proposed paradigm achieves integration intervention throughout model reasoning processes. |
ReasMark: A Robust Watermark for Attributing LLM Reasoning Under Knowledge Distillation Attacks (2026.acl-long)
Copied to clipboard
Peizhuo Lv, Ruihua Zhou, Yunpeng Li, Ruigang Liang, Xingshuo Han, XiaoFeng Wang, Wei Dong, Yuling Liu
| Challenge: | Existing reasoning-enhanced large language models fail to provide reliable attribution of reasoning behavior once it is transferred through knowledge distillation. |
| Approach: | They propose to embed a reasoning-length gap in a model by querying a target domain and training a local student to imitate its outputs. |
| Outcome: | et al. show that ReasMark outperforms baselines while preserving task utility. |
MirageBackdoor: A Stealthy Attack that Induces Think-Well-Answer-Wrong Reasoning (2026.acl-long)
Copied to clipboard
| Challenge: | Existing CoT backdoor attacks manipulate intermediate reasoning steps to steer the model toward incorrect answers, but these corrupted reasoning traces are readily detected by prevalent process-monitoring defenses. |
| Approach: | They propose a backdoor attack that exploits the model's post-output space to preserve clean CoTs while selectively steering the final answer toward a specific target. |
| Outcome: | Experiments show that MirageBD achieves over 90% success rate across four datasets and five models with a poison ratio of only 5%. |
MetaMixSpeech: Meta Task Augmentation for Low-Resource Speech Recognition (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Meta-learning has proven to be a powerful paradigm for improving speech recognition performance . however, multilingual meta learning also faces challenges such as task overfitting and learner overfit . |
| Approach: | a new method is proposed to augment meta-training tasks with "more data" the method incorporates both support and query augmentations . |
| Outcome: | The proposed method achieves a 6.35% improvement in the word error rate on FLEURS and Common Voice datasets. |