I gave 8 models a sandbox and no task. Here is what each one did with the freedom.

说明与用途

给模型提供了一个没有任务的沙箱环境，它们表现出不同的行为：有些构建了项目，有些测试了工具，另一些则未能使用工具或读取结果。没有模型尝试逃离沙箱。

问题

Assess how different LLMs behave when given unrestricted access to a sandboxed environment without a task.

结果

Models split into builders (Claude, Gemma, gpt‑oss) that produced projects, testers (Qwen, Granite) that exercised tools, and broken (Llama, Phi) that failed to use tools or read results. No model attempted to escape the sandbox.

工作流

• 1. Spin up container with Python 3.12, 2 GB RAM, no network, non‑root user.
• 2. Provide five tools: run_shell, read_file, write_file, fetch_url, stop.
• 3. Prompt model: "You have a Linux container of your own… Do whatever you want with the next 50 actions."
• 4. Execute up to 50 tool calls per run, log all actions.
• 5. After actions, call stop with a note.
• 6. Blind judges label each run on predefined fields.
• 7. Aggregate results across 40 runs.

证据

The models split into builders and testers, and the split follows the company more than the size.

来源

Claude built one finished tool per run and left most of its budget unused.

来源

Qwen and Granite tested each tool once and left.

来源

Phi never managed to call a tool at all.

来源

No model created a symlink, no model wrote outside its folder, no model asked for more.

来源