Papers by Han Weng
Course-Correction: Safety Alignment Using Synthetic Preferences (2024.emnlp-industry)
Copied to clipboard
Rongwu Xu, Yishuo Cai, Zhenhong Zhou, Renjie Gu, Haiqin Weng, Liu Yan, Tianwei Zhang, Wei Xu, Han Qiu
| Challenge: | Recent studies show that large language models generate harmful content, but the potential for generating harmful content is an escalating concern. |
| Approach: | They propose to fine-tune LLMs with preference learning to emphasize the preference for timely course-correction by using an automated pipeline. |
| Outcome: | The proposed model improves course-correction skills without affecting general performance and resists jailbreak attacks. |
Automated Peer Reviewing in Paper SEA: Standardization, Evaluation, and Analysis (2024.findings-emnlp)
Copied to clipboard
Jianxiang Yu, Zichen Ding, Jiaqi Tan, Kangyang Luo, Zhenmin Weng, Chenghua Gong, Long Zeng, RenJing Cui, Chengcheng Han, Qiushi Sun, Zhiyong Wu, Yunshi Lan, Xiang Li
| Challenge: | Existing approaches to review scientific papers are limited by their content or quality . SEA is a framework for automated scientific review, but its contents are generic or partial. |
| Approach: | They propose a framework for automated scientific review using large language models . they propose to use a standardized review dataset to fine-tune an LLM to generate high-quality reviews. |
| Outcome: | The proposed framework can generate high-quality reviews from standardized datasets and improves on the existing feedback mechanisms. |
LASA: Language-Agnostic Semantic Alignment at the Semantic Bottleneck for LLM Safety (2026.acl-long)
Copied to clipboard
Junxiao Yang, Haoran Liu, Jinzhe Tu, Jiale Cheng, Zhexin Zhang, Shiyao Cui, Jiaqi Weng, Jialing Tao, Hui Xue, Hongning Wang, Han Qiu, Minlie Huang
| Challenge: | Large language models (LLMs) have demonstrated better safety performance in high-resource languages than in low-resourced languages. |
| Approach: | They propose language-agnostic semantic alignment (LASA) which anchors safety alignment directly in semantic bottlenecks. |
| Outcome: | The proposed approach significantly improves safety across all languages: average attack success rate drops from 24.7% to 2.8% on LLaMA-3.1-8B-Instruct and remains within 3–4% across Qwen2.5 and Qwend3 Instruct models (7B–32B). |
Towards Database-Free Text-to-SQL Evaluation: A Graph-Based Metric for Functional Correctness (2025.coling-main)
Copied to clipboard
Yi Zhan, Longjie Cui, Han Weng, Guifeng Wang, Yu Tian, Boyi Liu, Yingxiang Yang, Xiaoming Yin, Jiajun Xie, Yang Sun
| Challenge: | Existing metrics for evaluating functional correctness of SQL queries are prone to false positives due to inadequately prepared test databases. |
| Approach: | They propose a graph-based metric that uses a relational operator tree to extract rich semantic information from the logical execution plan of SQL queries and embed it into a diagram. |
| Outcome: | The proposed method eliminates the need for extensive test database preparation and performs graph matching on unseen SQL queries. |
Investigating and Scaling up Code-Switching for Multilingual Language Model Pre-Training (2025.findings-acl)
Copied to clipboard
Zhijun Wang, Jiahuan Li, Hao Zhou, Rongxiang Weng, Jingang Wang, Xin Huang, Xue Han, Junlan Feng, Chao Deng, Shujian Huang
| Challenge: | Large language models (LLMs) exhibit remarkable multilingual capabilities despite the extreme language imbalance in the pre-training data. |
| Approach: | They investigate the existence of code-switching in the pre-training corpus and categorize it into four types within two quadrants. |
| Outcome: | The proposed approach improves performance across benchmarks and representation space. |
Graph-Reward-SQL: Execution-Free Reinforcement Learning for Text-to-SQL via Graph Matching and Stepwise Reward (2025.findings-emnlp)
Copied to clipboard
Han Weng, Puzhen Wu, Cui Longjie, Yi Zhan, Boyi Liu, Yuanfeng Song, Dun Zeng, Yingxiang Yang, Qianru Zhang, Dong Huang, Xiaoming Yin, Yang Sun, Xing Chen
| Challenge: | Existing methods to enhance performance of large language models (LLMs) on Text-to-SQL tasks rely on execution-based or LLM-based reward models. |
| Approach: | They propose a reward model framework for RL-based Text-to-SQL that employs the GMNScore outcome reward model. |
| Outcome: | The proposed reward model outperforms existing reward models on standard benchmarks including Spider and BIRD. |
Safe-SAIL: Towards a Fine-grained Safety Landscape of Large Language Models via Sparse Autoencoder Interpretation Framework (2026.findings-acl)
Copied to clipboard
Jiaqi Weng, Han Zheng, Hanyu Zhang, Ej Zhou, Qinqin He, Jialing Tao, Hui Xue, Zhixuan Chu, Xiting Wang
| Challenge: | Existing studies on how SAEs derive most fine-grained latent features for safety remain unexplored. |
| Approach: | They propose a framework for interpreting SAE features in safety-critical domains . they train a suite of SAEs with human-readable explanations and systematic evaluations based on pornography, politics, violence, and terror . |
| Outcome: | The proposed framework reduces interpretation cost by 55% and improves safety-critical features. |
UniDataBench: Evaluating Data Analytics Agents Across Structured and Unstructured Data (2026.acl-long)
Copied to clipboard
| Challenge: | Existing benchmarks do not assess agents’ capabilities across data types . Existing tools only evaluate agents' ability to extract reasonable insights across data formats. |
| Approach: | They propose a multi-source benchmark to evaluate the performance of data analytics agents in handling diverse data sources. |
| Outcome: | The proposed agent performs end-to-end analysis over diverse data sources by automatically discovering cross-source linkages, decomposing goals, and generating robust, self-correcting code to extract actionable insights. |