Pengyue Jia
Pengyue Jia has 3 papers in our library, in 2025–2026, most of them on Language models & pretraining, Reasoning, Bias, fairness & safety; often with Maolin Wang, Xiangyu Zhao, Jingyu Peng.
OpenAlex · ORCID · Wikidata · h-index 6 · 62 works · 139 citations OpenAlex, 2026-10-14
📄 3 works in our library · 🔗 Cited 15 times in our library · 1 more in our library without a page yet · ⭐ Paper Stars ranking: #9133
Affiliations
Fields, topics and concepts
📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Arts and Humanities · 3 worksComputer Science · 3 worksArtificial Intelligence · 3 worksLanguage and Linguistics · 3 worksInformation Systems · 1 work
🧭 Topics 3Research subjects and languages the work is about.Language models & pretraining · 3 worksReasoning · 2 worksBias, fairness & safety · 1 work
Works by year
2020s
Papers
- Stepwise Reasoning Disruption Attack of LLMs (2025 · 10 citations)
LLMs are often accessed through third-party platforms and APIs. This means an intermediary could quietly tamper with a model's reasoning while the output still looks normal. The authors study this risk for multi-step reasoning, where one flawed early step can spread through the whole chain. Earlier reasoning attacks… - Logic Jailbreak: Efficiently Unlocking LLM Safety Restrictions Through Formal Logical Expression (2026 · 8 citations)
The paper asks why safety-aligned LLMs still give in to jailbreaks. The authors propose that alignment learns refusal mostly from surface token patterns, so a harmful request with the same meaning but an unfamiliar form can slip past safeguards. To test this, they present LogiBreak, a black-box attack that rewrites a… - MemSearch-o1: Empowering Large Language Models with Reasoning-Aligned Memory Growth in Agentic Search (2026 · 1 citation)
Agentic (deep) search lets an LLM alternate between reasoning and retrieval. Each round adds thinking traces and retrieved passages, so the context grows and key evidence gets buried. The authors call this memory dilution. Existing fixes that summarize, filter or prune memory often miss the shifting intent of…
All works
- Stepwise Reasoning Disruption Attack of LLMsAnnual Meeting of the Association for Computational Linguistics (ACL)Jingyu Peng, Maolin Wang, Xiangyu Zhao, et al.2025 · 10 citations
- Logic Jailbreak: Efficiently Unlocking LLM Safety Restrictions Through Formal Logical ExpressionFindings of the Association for Computational LinguisticsJingyu Peng, Maolin Wang, Nan Wang, et al.2026 · 8 citations
- MemSearch-o1: Empowering Large Language Models with Reasoning-Aligned Memory Growth in Agentic SearchAnnual Meeting of the Association for Computational Linguistics (ACL)Sheng Zhang, Junyi Li, Yingyi Zhang, et al.2026 · 1 citation
- Learning How and What to Memorize: Cognition-Inspired Two-Stage Optimization for Evolving Memory ↗Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · no summary yetDerong Xu et al.2026
Co-authors here
- Maolin Wang ×3
- Xiangyu Zhao ×3
- Jingyu Peng ×2
- Kai Zhang ×2
- Wanyu Wang ×2
- Jiatong Li
- Junyi Li
- Nan Wang
- Qi Liu
- Sheng Zhang
- Yong Liu
- Yuchen Li
Venues
- Annual Meeting of the Association for Computational Linguistics (ACL) ×3: ACL 2026, ACL 2025
- Findings of the Association for Computational Linguistics: Findings of the Association for Computational Linguistics 2026
Cited by
Updated
This page in Markdown → https://md.research.elamur.ai/free/en/authors/pengyue-jia/