Papers by Xiaofei Zhou
SDC-LoRA: Singular-Subspace Drift Controlled LoRA to Mitigate Knowledge Forgetting (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to adapt LLMs to new tasks focus on limiting knowledge forgetting . et al., 2023b) suggest a solution to this problem by limiting update energy in the principal singular subspace of W0 . |
| Approach: | They propose a low-rank Adaptation (LoRA) that steers early updates away from principal directions and mitigates forgetting by constraining update energy in the principal singular subspace of W0. |
| Outcome: | The proposed model mitigates forgetting on MMLU, TruthfulQA, and HellaSwag while keeping minor-subspace updates unchanged. |
RedCoder: Automated Multi-Turn Red Teaming for Code LLMs (2026.acl-long)
Copied to clipboard
Wenjie Jacky Mo, Qin Liu, Xiaofei Wen, Dongwon Jung, Hadi Askari, Wenxuan Zhou, Zhe Zhao, Muhao Chen
| Challenge: | Existing red-teaming approaches for code generation rely on extensive human effort and are prone to generating malicious code under adversarial environments. |
| Approach: | They propose a red-teaming agent that engages victim models in multi-turn conversations to elicit vulnerable code. |
| Outcome: | Experiments show that RedCoder outperforms red-teaming methods in inducing vulnerabilities in code generation. |
Exploiting Emotion-Semantic Correlations for Empathetic Response Generation (2023.findings-emnlp)
Copied to clipboard
Zhou Yang, Zhaochun Ren, Wang Yufeng, Xiaofei Zhu, Zhihao Chen, Tiecheng Cai, Wu Yunbing, Yisong Su, Sibo Ju, Xiangwen Liao
| Challenge: | Empathetic response generation aims to generate empathetic responses by understanding the speaker’s emotional feelings from the language of dialogue. |
| Approach: | They propose a dynamical Emotion-Semantic Correlation Model (ESCM) which constructs dynamic emotion-semantics through the interaction of context and emotions. |
| Outcome: | The proposed model understands emotions more accurately and expresses fluent and informative empathetic responses. |
Learning Dialogue Representations from Consecutive Utterances (2022.naacl-main)
Copied to clipboard
| Challenge: | Dialogue Sentence Embedding (DSE) is a self-supervised contrastive learning method that learns effective dialogue representations suitable for a wide range of dialogue-oriented tasks. |
| Approach: | They propose a self-supervised contrastive learning method that learns dialogue representations suitable for a wide range of dialogue tasks. |
| Outcome: | The proposed method outperforms baselines on five dialogue tasks on a few-shot and zero-shot datasets. |
ThinkGuard: Deliberative Slow Thinking Leads to Cautious Guardrails (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing guardrails rely on rule-based filtering or single-pass classification, limiting their ability to handle nuanced safety violations. |
| Approach: | They propose a critique-augmented guardrail model that distills knowledge from high-capacity LLMs by generating structured critiques alongside safety labels. |
| Outcome: | The proposed model outperforms existing guardrail models on multiple safety benchmarks and achieves the highest average F1 and AUPRC. |
Discourse Marker Augmented Network with Reinforcement Learning for Natural Language Inference (P18-1)
Copied to clipboard
| Challenge: | Existing approaches to natural language inference focus on interaction architectures of sentences . but, we propose to transfer knowledge from discourse markers to augment the model . |
| Approach: | They propose to transfer knowledge from discourse markers to augment the quality of the NLI model. |
| Outcome: | The proposed method achieves state-of-the-art performance on large-scale datasets. |
Jailbreak Large Vision-Language Models Through Multi-Modal Linkage (2025.acl-long)
Copied to clipboard
| Challenge: | Existing methods to jailbreak large vision-language models fail against cutting-edge models such as GPT-4o, despite having undergone safety alignment training. |
| Approach: | They propose a new framework for jailbreaking large vision-language models that uses an encryption-decryption process to mitigate the over-exposure of harmful information. |
| Outcome: | The proposed framework jailbreaks GPT-4o with 99.40% success rates on SafeBench, 98.81% on MM-SafeBench and 99.07% on HADES-Dataset. |
Temporal Evidence Chain for Temporal Knowledge Graph Question Answering with Large Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Temporal Knowledge Graph Question Answering (TKGQA) aims to answer temporal questions using knowledge from Temporal knowledge graphs. |
| Approach: | They propose a framework to construct temporal evidence chains for LLM reasoning using Temporal Knowledge Graphs. |
| Outcome: | TECQA outperforms existing methods on MultiTQ and CronQuestions. |
An Iterative Associative Memory Model for Empathetic Response Generation (2024.acl-long)
Copied to clipboard
| Challenge: | Existing methods for empathetic response generation ignore the associated words between dialogue utterances. |
| Approach: | They propose an iterative associative memory model to capture associated words between dialogue utterances and situations, dialogue history, and a memory module for storing associated words. |
| Outcome: | The proposed model captures key words between dialogue utterances and situations, dialogue history, and a memory module, thereby accurately and nuancedly comprehending the utterables. |