When Benign Inputs Lead to Severe Harms: Eliciting Unsafe Unintended Behaviors of Computer-Use Agents
Paper
•
2602.08235
•
Published
AI & ML interests
Natural language processing, language models, language agents
Recent Activity
View all activity
Papers
When Actions Go Off-Task: Detecting and Correcting Misaligned Actions in Computer-Use Agents
When Benign Inputs Lead to Severe Harms: Eliciting Unsafe Unintended Behaviors of Computer-Use Agents
nnnyt
authored
a
paper
2 days ago
Add image-text-to-text task category to metadata
#1 opened 9 days ago
by
nielsr
WeijianQi1999
updated
a
Space
7 days ago
nnnyt
authored
a
paper
9 days ago
nnnyt
submitted
a
paper
to Daily Papers
10 days ago
xuetianci99
updated
a
collection
11 days ago
xuetianci99
updated
a
model
11 days ago
xuetianci99
published
a
model
11 days ago
xuetianci99
updated
a
model
11 days ago
xuetianci99
published
a
model
11 days ago
xuetianci99
updated
a
model
11 days ago
xuetianci99
published
a
model
11 days ago
xuetianci99
updated
a
model
11 days ago
xuetianci99
published
a
model
11 days ago