1 citations · 1 across the 10 of their papers we have counts for
4 papers · 2 filters
DrawingBench: Evaluating Spatial Reasoning and UI Interaction Capabilities of Large Language Models through Mouse-Based Drawing Tasks
Hyunjun Kim, Sooyoung Ryu
As agentic AI systems increasingly operate autonomously, establishing trust through verifiable evaluation becomes critical. Yet existing benchmarks lack the transparency and audita…
X-Teaming Evolutionary M2S: Automated Discovery of Multi-turn to Single-turn Jailbreak Templates
Hyunjun Kim, Junwoo Ha, Sangyoon Yu +1
Multi-turn-to-single-turn (M2S) compresses iterative red-teaming into one structured prompt, but prior work relied on a handful of manually written templates. We present X-Teaming…
ObjexMT: Objective Extraction and Metacognitive Calibration for LLM-as-a-Judge under Multi-Turn Jailbreaks
Hyunjun Kim, Junwoo Ha, Sangyoon Yu +1
LLM-as-a-Judge (LLMaaJ) enables scalable evaluation, yet we lack a decisive test of a judge's qualification: can it recover the hidden objective of a conversation and know when tha…
M2S: Multi-turn to Single-turn jailbreak in Red Teaming for LLMs
Junwoo Ha, Hyunjun Kim, Sangyoon Yu +4
We introduce a novel framework for consolidating multi-turn adversarial ``jailbreak'' prompts into single-turn queries, significantly reducing the manual overhead required for adve…