Papers by Yumin Kim
Keep Security! Benchmarking Security Policy Preservation in Large Language Model Contexts Against Indirect Attacks in Question Answering (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are increasingly deployed in sensitive domains . large-scale benchmarks for contextual security preservation against attacks remain lacking . |
| Approach: | They evaluate 10 Large Language Models on a benchmark dataset to assess their adherence to contextual non-disclosure policies. |
| Outcome: | The proposed model fails to adhere to user-defined security policies in question answering . the model fails in indirect attacks, especially when it violates user-definable policies . |
KoCoSa: Korean Context-aware Sarcasm Detection Dataset (2024.lrec-main)
Copied to clipboard
| Challenge: | Sarcasm is a form of verbal irony where someone says the opposite of what they mean . misunderstanding this sarcasm may lead to fatal errors in dialogue systems . |
| Approach: | They propose a dataset for the Korean dialogue sarcasm detection task that uses 12.8K daily Korean dialogues and the labels on the last response. |
| Outcome: | The proposed system outperforms strong baselines like large language models in the Korean sarcasm detection task. |
Personality Editing for Language Models through Adjusting Self-Referential Queries (2026.eacl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are integral to conversational agents and content creation, but they lack robustness and require large-scale training data to achieve significant improvements in personality alignment. |
| Approach: | They propose a method that introduces adjustment queries where self-referential statements grounded in psychological constructs are treated analogously to factual knowledge to enable direct editing of personality-related responses. |
| Outcome: | The proposed method improves personality alignment across personality dimensions and requires only 12 editing samples to achieve significant improvements. |