Story
arxiv_cs_ai · Nov 1, 2026 · paper
Source brief
Is Your Long-Horizon Benchmark Really Deep? Measuring and Synthesizing Depth in Terminal Tasks
arxiv.orgNov 1, 2026
original source linked
In brief
Long-horizon terminal tasks are used to evaluate and train agents. Human-curated datasets are scarce, so many synthesis methods have emerged. They usually measure the length of a task by the turns a model needs to pas...
Feed lens
agenteval