Papers by Yubin Choi
Don’t Miss the Forest for the Trees: Attentional Vision Calibration for Large Vision Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Large Vision Language Models suffer from hallucinations, attributing incorrect or misleading features to images. |
| Approach: | They propose a test-time approach that recalibrates the influence of blind tokens . they identify blind token by analyzing layer-wise attention distributions over image tokens. |
| Outcome: | The proposed approach reduces hallucinations in large vision language models . it uses a contrastive decoding strategy to balance the influence of blind tokens . |
FinHarmBench: Financial Jailbreak Benchmark and Unsupervised Safety Fine-Tuning via Refusal Steering Distillation (2026.acl-industry)
Copied to clipboard
Yubin Choi, Yujin Yang, Subin Kim, Seokil Ham, Seungju Cho, Jungmin Son, Youngjun Kwak, Changick Kim
| Challenge: | Existing safety benchmarks focus on general harms and lack the granularity needed to capture domain-specific financial threats. |
| Approach: | They propose a benchmark to evaluate financially harmful and confusable benign prompts. |
| Outcome: | The proposed framework improves refusal behavior without annotating refusal responses. |
Detecting Offensive Language in an Open Chatbot Platform (2024.lrec-main)
Copied to clipboard
Hyeonho Song, Jisu Hong, Chani Jung, Hyojin Chin, Mingi Shin, Yubin Choi, Junghoi Choi, Meeyoung Cha
| Challenge: | Existing efforts to automatically filter offensive language are vulnerable to users’ deliberate text manipulation tactics, such as misspelling words. |
| Approach: | They propose a contrastive learning model that embeds chat content with a random masking strategy to detect offensive language in open-domain chat conversations. |
| Outcome: | The proposed model outperforms existing models in detecting offensive language in open-domain chat conversations while also showing robustness against users’ deliberate text manipulation tactics when using offensive language. |