@article{cai2026realcontrolbench,
title = {{RealControlBench}: A Benchmark for Interpretability-Informed Fine-Grained Control of {LLM}s and Agents},
author = {Cai, Xinying and Hu, Bin and Jin, Mingyu and Zhang, Qiming and Ouyang, Yang and Guo, Bangwei and Guo, Minghao and Yin, Yutong and Gao, Lang and Han, Feijiang and Yu, Sheldon and Liang, Kaiqu and Zeng, Qingcheng and Niu, Jingcheng and Yang, Fan and Cheng, Wei and Xiang, Zhen and McAuley, Julian and Wang, Zhaoran},
journal = {arXiv preprint arXiv:2610.27295},
year = {2026},
}Publications
- 2026RealControlBench: A Benchmark for Interpretability-Informed Fine-Grained Control of LLMs and AgentsarXiv preprintXinying Cai, Bin Hu, Mingyu Jin, Qiming Zhang, Yang Ouyang, Bangwei Guo, Minghao Guo, Yutong Yin, Lang Gao, Feijiang Han, Sheldon Yu, Kaiqu Liang, Qingcheng Zeng, Jingcheng Niu, Fan Yang, Wei Cheng, Zhen Xiang, Julian McAuley and Zhaoran Wang.
- Probe with Participation Trophies: Random-Reward RL as a Probe of LLM CapabilityarXiv preprintYu Mao, Lei Yu, Zining Zhu, Yusheng Zheng, Haohang Li, Freda Shi, Yutong Yin, Zhaoran Wang and Jingcheng Niu.
- Agent MechSuits: Mechanistic Subspace Safety Steering for Multi-Turn CLI AgentsNeurIPS 2026Weidi Luo, Qiming Zhang, Yihao Quan, Mingyu Jin, Jie Cai, Chaowei Xiao, Jingcheng Niu and Zhen Xiang.
- Farther the Shift, Sparser the Representation: Analyzing OOD Mechanisms in LLMsNeurIPS 2026Mingyu Jin, Yutong Yin, Jingcheng Niu†, Qingcheng Zeng, Wujiang Xu, Mengnan Du, Wei Cheng, Zhaoran Wang, Tianlong Chen and Dimitris N. Metaxas.
- Many Circuits, One Mechanism: Input Variation and Evaluation Granularity in Circuit DiscoveryTMLR 2026 Featured Certification
- All Circuits Lead to Rome: Rethinking Functional Anisotropy in Circuit and Sheaf Discovery for LLMsICML 2026Xi Chen*, Mingyu Jin*, Jingcheng Niu*, Yutong Yin, Jinman Zhao, Bangwei Guo, Dimitris N. Metaxas, Zhaoran Wang, Yutao Yue and Gerald Penn.
- Hypothesis-Driven Feature Manifold Analysis in LLMs via Supervised Multi-Dimensional ScalingTMLR 2026
- 2025Illusion or Algorithm? Investigating Memorization, Emergence, and Symbolic Processing in In-Context LearningTMLR 2025 J2C Certification
- Sheaf Discovery with Joint Computation Graph Pruning and Flexible GranularityEMNLP 2025
- Tiny Budgets, Big Gains: Parameter Placement Strategy in Parameter Super-Efficient Fine-TuningEMNLP 2025
- Llama See, Llama Do: A Mechanistic Perspective on Contextual Entrainment and Distraction in LLMsACL 2025 Outstanding Paper Award
- Attenuating Catastrophic Forgetting of Token-Level Knowledge in Contrastive Sentence Embeddings with Attentive PoolingCAIAC 2025
- 2024ConTempo: A Unified Temporally Contrastive Framework for Temporal Relation ExtractionACL 2024 Findings
- What does the Knowledge Neuron Thesis Have to do with Knowledge?ICLR 2024 spotlight
- 2023Discourse Information for Document-Level Temporal Dependency ParsingCODI 2023 @ ACL
- 2022Using Roark-Hollingshead Distance to Probe BERT's Syntactic CompetenceBlackboxNLP 2022 @ EMNLP
- Does BERT Rediscover a Classical NLP Pipeline?COLING 2022 oral
- Chinese Quantifier Scope, Concord, and Lexical Resource SemanticsHPSG 2022
- 2021Statistically Evaluating Social Media Sentiment Trends towards COVID-19 Non-Pharmaceutical Interventions with Event StudiesSMM4H 2021 @ NAACL
- 2020Grammaticality and Language ModellingEval4NLP 2020 @ EMNLPJingcheng Niu and Gerald Penn.
- Temporal Histories of Epidemic Events (THEE): A Case Study in Temporal Annotation for Public HealthLREC 2020
- Application of natural language processing algorithms for extracting information from news articles in event-based surveillanceCCDR 2020Victoria Ng, Erin E. Rees, Jingcheng Niu, Abdelhamid Zaghlool, Homeira Ghiasbeglou and Adrian Verster.
- 2019Rationally Reappraising ATIS-based Dialogue SystemsACL 2019Jingcheng Niu and Gerald Penn.