Picture for Kimin Lee

Kimin Lee

Silent Branding Attack: Trigger-free Data Poisoning Attack on Text-to-Image Diffusion Models

Add code
Mar 12, 2025
Viaarxiv icon

Subtask-Aware Visual Reward Learning from Segmented Demonstrations

Add code
Feb 28, 2025
Viaarxiv icon

DiffExp: Efficient Exploration in Reward Fine-tuning for Text-to-Image Diffusion Models

Add code
Feb 19, 2025
Viaarxiv icon

Understanding Impact of Human Feedback via Influence Functions

Add code
Jan 10, 2025
Viaarxiv icon

Improbable Bigrams Expose Vulnerabilities of Incomplete Tokens in Byte-Level Tokenizers

Add code
Oct 31, 2024
Figure 1 for Improbable Bigrams Expose Vulnerabilities of Incomplete Tokens in Byte-Level Tokenizers
Figure 2 for Improbable Bigrams Expose Vulnerabilities of Incomplete Tokens in Byte-Level Tokenizers
Figure 3 for Improbable Bigrams Expose Vulnerabilities of Incomplete Tokens in Byte-Level Tokenizers
Figure 4 for Improbable Bigrams Expose Vulnerabilities of Incomplete Tokens in Byte-Level Tokenizers
Viaarxiv icon

MobileSafetyBench: Evaluating Safety of Autonomous Agents in Mobile Device Control

Add code
Oct 23, 2024
Figure 1 for MobileSafetyBench: Evaluating Safety of Autonomous Agents in Mobile Device Control
Figure 2 for MobileSafetyBench: Evaluating Safety of Autonomous Agents in Mobile Device Control
Figure 3 for MobileSafetyBench: Evaluating Safety of Autonomous Agents in Mobile Device Control
Figure 4 for MobileSafetyBench: Evaluating Safety of Autonomous Agents in Mobile Device Control
Viaarxiv icon

When LLMs Go Online: The Emerging Threat of Web-Enabled LLMs

Add code
Oct 18, 2024
Viaarxiv icon

Latent Action Pretraining from Videos

Add code
Oct 15, 2024
Figure 1 for Latent Action Pretraining from Videos
Figure 2 for Latent Action Pretraining from Videos
Figure 3 for Latent Action Pretraining from Videos
Figure 4 for Latent Action Pretraining from Videos
Viaarxiv icon

Automated Filtering of Human Feedback Data for Aligning Text-to-Image Diffusion Models

Add code
Oct 14, 2024
Figure 1 for Automated Filtering of Human Feedback Data for Aligning Text-to-Image Diffusion Models
Figure 2 for Automated Filtering of Human Feedback Data for Aligning Text-to-Image Diffusion Models
Figure 3 for Automated Filtering of Human Feedback Data for Aligning Text-to-Image Diffusion Models
Figure 4 for Automated Filtering of Human Feedback Data for Aligning Text-to-Image Diffusion Models
Viaarxiv icon

DiffusionGuard: A Robust Defense Against Malicious Diffusion-based Image Editing

Add code
Oct 08, 2024
Viaarxiv icon