Papers by Jeesu Jung
Guidance-Based Prompt Data Augmentation in Specialized Domains for Named Entity Recognition (2024.acl-short)
Copied to clipboard
| Challenge: | specialized fields such as science and biology face significant challenges due to the scarcity of quality data. |
| Approach: | They propose a guidance data augmentation technique that abstracts context and sentence structure and maintains context-entity relationships for DA. |
| Outcome: | The proposed method enhances the training performance of named entity recognition tasks while maintaining context-entity relationships. |
ZEBRA: Leveraging Model-Behavioral Knowledge for Zero-Annotation Preference Dataset Construction (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Recent efforts in LLM alignment focus on instance-wise supervision, costing substantial . ZEBRA binarizes response pairs by evaluating the quality and similarity of their origin models . |
| Approach: | They propose a model behavior-wise zero-annotation framework that binarizes preference data . ZEBRA binarized response pairs by evaluating the quality and similarity of their origin models . |
| Outcome: | The proposed framework achieves comparable alignment performance to instance-supervised methods . |
Tracing Logit Trajectories Across Layer Depth: Dataset-Level Explainability for Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Sentence-level explanations miss the bigger picture of how a black-box model behaves across data . a dataset-level analysis that traces the intermediate structure of decision formation is needed . |
| Approach: | They propose a method that aggregates logit updates into a reproducible dataset-level trajectory pattern. |
| Outcome: | The proposed model enables depth-wise explainability across 6 languages and 5 NLP tasks. |
Exploring Domain Robust Lightweight Reward Models based on Router Mechanism (2024.findings-acl)
Copied to clipboard
| Challenge: | Recent advances in large language models have relied on the large reward model for fine-tuning, but the use of a single reward model across domains may not always be optimal. |
| Approach: | They propose to use router mechanisms to train small language models in a domain-specific manner . they use internal routers, external routers and router adapters to create a single reward model . |
| Outcome: | The proposed approach reduces parameter size while minimizing parameter size. |
AMACE: Automatic Multi-Agent Chart Evolution for Iteratively Tailored Chart Generation (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent studies have shown that chart generation requires manual input of code and intent . despite the benefits of large language models, chart generation still requires user input of many parameters . |
| Approach: | They propose a loop-based framework for automatically evolving charts in a multi-agent environment using large language models. |
| Outcome: | The proposed framework improves performance by 29.97% compared to first generation while reducing generation time by 86.9% compared with manual prompt-based methods. |
Courtroom-LLM: A Legal-Inspired Multi-LLM Framework for Resolving Ambiguous Text Classifications (2025.coling-main)
Copied to clipboard
| Challenge: | Using a multi-LLM structure inspired by legal courtroom processes, we demonstrate that it can improve decision-making accuracy in ambiguous text classification scenarios. |
| Approach: | They propose a legal-inspired multi-LLM structure that simulates a courtroom setting within LLMs and assigns roles similar to those of prosecutors, defense attorneys, and judges. |
| Outcome: | The proposed model outperforms both single-LLM classifiers and simpler multi-LLMS setups in ambiguous text classification tasks. |
When Does Auxiliary Modality Matter in Solving Geometric Problems? A Comprehensive Study of Textual, Formal, and Visual Modalities (2026.eacl-short)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) face challenges in integrating linguistic and spatial reasoning, which limits their performance on geometry problems. |
| Approach: | They compare four auxiliary modalities on open- and closed-source multimodal LLMs . they show that DES boosts the accuracy of open-source LLM models . |
| Outcome: | The proposed modalities improve performance on open- and closed-source LLMs. |
Empirical Analysis of Task Mixture Effects in Small-scale Instruction Tuning: A Statistical Approach (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent studies have shown that instruction tuning can significantly vary depending on the task and format diversity of training data. |
| Approach: | They conduct experiments on unlabeled instruction corpora to identify effective mixtures. |
| Outcome: | The proposed model can achieve up to 5.7 speedup in training with 1,000 curated examples. |