Clarify or Answer: Reinforcement Learning for Agentic VQA with Context Under-specification
Reinforcement learning for agentic VQA that balances clarification and answering under underspecified context.
•
1 min read
Read more
Reinforcement learning for agentic VQA that balances clarification and answering under underspecified context.
Benchmarking dark pattern susceptibility of computer-use agents in realistic UI environments.