Papers2

#long-context agents

LOCA-bench: Benchmarking Language Agents Under Controllable and Extreme Context Growth

Weihao Zeng, Yuzhen Huang et al.Feb 8arXiv

LOCA-bench is a test that challenges AI agents to work correctly as their to-do list and background information grow very, very long.

#LOCA-bench#long-context agents#context rot

AgentLongBench: A Controllable Long Benchmark For Long-Contexts Agents via Environment Rollouts

Intermediate

Shicheng Fang, Yuxin Wang et al.Jan 28arXiv

AgentLongBench is a new test that checks how well AI agents think over very long stories made of their own actions and the world's replies, not just by reading static documents.

#AgentLongBench#long-context agents#environment rollouts