Skip to content

Dataset · camel-ai

Terminal RL Agent Training Environments (SETA-Env)

camel-ai/SETA-Env

4,567 verifiable terminal environments for reinforcement learning and autonomous AI agent evaluation from CAMEL-AI.

What's inside

The release from CAMEL-AI consists of 4,567 self-contained task directories in Harbor format for running and evaluating English-language terminal agents. The dataset is split into two subsets: 3,255 synthesized tasks (SETA_Synth) and 1,312 evolved variations (SETA_Evolve).

What you can do with it

  • Training autonomous console agents via reinforcement learning
  • Benchmarking and testing LLM capability to solve terminal tasks
Tasks
Reinforcement learning
Languages
English
Size
1K–10K rows
Downloads this month
263.2K
Likes
17

How to load it

from datasets import load_dataset

ds = load_dataset("camel-ai/SETA-Env")