4 papers
VGI-Bench: Probing Visual Intelligence in Video Generation Models
Xuan He, Cong Wei, Yuhao Cheng +20
Recent studies suggest that video generation models can exhibit certain forms of zero-shot visual reasoning through generated frames. Yet reliable evaluation remains challenging: b…
RewardHarness: Self-Evolving Agentic Post-Training
Yuxuan Zhang, Penghui Du, Bo Li +11
Aligning generative models with human preferences remains a central challenge. While RLHF is widely adopted, it relies on large-scale preference data. This limitation is particular…
ClawBench: Can AI Agents Complete Everyday Online Tasks?
Yuxuan Zhang, Yubo Wang, Yipeng Zhu +25
AI agents may be able to assist with emails and documents, but can they reliably complete everyday online workflows on real websites? Everyday online tasks offer a realistic yet un…
Under the Influence: Evaluating and Improving LLM Vigilance in Agentic Tasks
Sasha Robinson, Katherine M. Collins, Ilia Sucholutsky +1
Large Language Models (LLMs) are becoming increasingly autonomous assistants, capable of accomplishing series of complex tasks by interacting with open world domains such as softwa…