Publications

  1. Figure for Rep2Text: Decoding Full Text from a Single LLM Token Representation NeurIPS
    Rep2Text: Decoding Full Text from a Single LLM Token Representation
    Haiyan Zhao, Zirui He, Fan Yang, Ali Payani, Mengnan Du
    Conference on Neural Information Processing Systems (NeurIPS), 2026

  2. Figure for HyperPathsRAG: Multi-Target Path Approximation over Hypergraph Knowledge Graphs for Multi-Hop QA EMNLP Findings
    HyperPathsRAG: Multi-Target Path Approximation over Hypergraph Knowledge Graphs for Multi-Hop QA
    Guanhua Zhu, Zirui He, Mengnan Du, Chenxi Yuan
    Findings of the Association for Computational Linguistics: EMNLP 2026

  3. Figure for Universal Activation Verbalizer: A Unified Framework for Cross-Model Activation Explanation EMNLP Main
    Universal Activation Verbalizer: A Unified Framework for Cross-Model Activation Explanation
    Haiyan Zhao, Zirui He, Guanchu Wang, Ali Payani, Yingcong Li, Mengnan Du
    Conference on Empirical Methods in Natural Language Processing (EMNLP), Main Conference, 2026

  4. Figure for SAIF: A Sparse Autoencoder Framework for Interpreting and Steering Instruction Following of Language Models ECML-PKDD
    SAIF: A Sparse Autoencoder Framework for Interpreting and Steering Instruction Following of Language Models
    Zirui He, Haiyan Zhao, Yiran Qiao, Fan Yang, Ali Payani, Jing Ma, Mengnan Du
    European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD), 2026

  5. Figure for SAE-FiRE: Enhancing Earnings Surprise Predictions Through Sparse Autoencoder Feature Selection ACL Findings
    SAE-FiRE: Enhancing Earnings Surprise Predictions Through Sparse Autoencoder Feature Selection
    Huopu Zhang, Yanguang Liu, Miao Zhang, Zirui He, Mengnan Du
    Findings of the Association for Computational Linguistics: ACL 2026

  6. Figure for SAE-SSV: Supervised Steering in Sparse Representation Spaces for Reliable Control of Language Models EMNLP
    SAE-SSV: Supervised Steering in Sparse Representation Spaces for Reliable Control of Language Models
    Zirui He, Mingyu Jin, Bo Shen, Ali Payani, Yongfeng Zhang, Mengnan Du
    Conference on Empirical Methods in Natural Language Processing (EMNLP), 2025

  7. Figure for Mitigating shortcuts in language models with soft label encoding COLING
    Mitigating shortcuts in language models with soft label encoding
    Zirui He, Huiqi Deng, Haiyan Zhao, Ninghao Liu, Mengnan Du
    International Conference on Computational Linguistics (COLING), 2024