Papers by Junyi Zhu
FastMem: Fast Memorization of Prompt Improves Context Awareness of Large Language Models (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large language models struggle with context awareness, leading to inaccuracies in tasks requiring faithful adherence to provided information. |
| Approach: | They propose a method to enhance LLMs' context awareness by updating only the last Feed-Forward Network module to maximize the likelihood of the prompt before inference . |
| Outcome: | The proposed method improves the accuracy of Llama 3-8B-Inst on the NQ-SWAP dataset from 59.1% to 71.6% and reduces the output structure failure rate of Qwen 1.5-4B-Chat from 34.9% to 25.5%. |
GrantRel: Grant Information Extraction via Joint Entity and Relation Extraction (2021.findings-acl)
Copied to clipboard
| Challenge: | a funder name refers to an agency, organization, or program providing financial support for the research. |
| Approach: | They propose a funding sentence classifier and a relation extraction framework to extract grant information from scientific articles. |
| Outcome: | The proposed framework outperforms state-of-the-art BERT-based RE baselines against the PubMed Central and arXiv test sets. |
PERM: Psychology-grounded Empathetic Reward Modeling for Large Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing reward models evaluate empathy from a single perspective, overlooking bidirectional interaction nature of empathy. |
| Approach: | They propose a reward model that evaluates empathy from a single perspective . they propose PERM to integrate a bystander perspective to monitor overall interaction quality . |
| Outcome: | a new reward model outperforms state-of-the-art models on an emotional intelligence benchmark and an industrial daily conversation dataset. |
Confidence-Calibrated Small-Large Language Model Collaboration for Cost-Efficient Reasoning (2026.eacl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) have superior reasoning capabilities compared to small language models, but incur substantially higher inference costs. |
| Approach: | They propose a system that cascades an LLM with an SLM to achieve a balance between accuracy and cost in complex reasoning tasks. |
| Outcome: | The proposed system improves the SLM’s reasoning ability and confidence calibration across diverse datasets and model backbones. |
Multi-Task Pre-Finetuning of Lightweight Transformer Encoders for Text Classification and NER (2025.emnlp-industry)
Copied to clipboard
| Challenge: | nave multitask pre-finetuning introduces conflicting optimization signals that degrade overall performance. |
| Approach: | They propose a framework that enables a single shared encoder backbone with modular adapters. |
| Outcome: | The proposed framework achieves comparable performance to individual pre-finetuning while meeting practical deployment constraint. |
Decoding Text Spans for Efficient and Accurate Named-Entity Recognition (2026.acl-industry)
Copied to clipboard
| Challenge: | Named Entity Recognition (NER) is a key component in industrial information extraction pipelines, where systems must satisfy strict latency and throughput constraints in addition to strong accuracy. |
| Approach: | They propose a span-based NER framework that can be used to compute span representations at the final transformer stage, avoiding redundant computation in earlier layers. |
| Outcome: | The proposed framework matches competitive baselines while improving throughput and reducing computational cost. |
Adversarial Preference Learning for Robust LLM Alignment (2025.findings-acl)
Copied to clipboard
Yuanfu Wang, Pengyu Wang, Chenyang Xi, Bo Tang, Junyi Zhu, Wenqiang Wei, Chen Chen, Chao Yang, Jingfeng Zhang, Chaochao Lu, Yijun Niu, Keming Mao, Zhiyu Li, Feiyu Xiong, Jie Hu, Mingchuan Yang
| Challenge: | Modern language models rely on Reinforcement Learning from Human Feedback (RLHF) to encourage safe behaviors, but they remain vulnerable to adversarial attacks due to three key limitations: (1) the inefficiency and high cost of human annotation; (2) the vast diversity of potential adversarials; and (3) the risk of feedback bias and reward hacking. |
| Approach: | They propose an iterative adversarial training method that incorporates three key innovations to address these challenges. |
| Outcome: | Experiments on Mistral-7B-Instruct-v0.3 show that the proposed method significantly enhances robustness and reduces harmful outputs from 5.88% to 0.43%. |