---
format: "aidr-story-markdown/v1"
id: "97564ba477f543718c25f71d84f8f41db1491348e6a87aff296442000ab93e31"
canonical_url: "https://aidr.today/97564ba4?lang=en"
title: "Causal Discovery Bench: Can LLMs Autonomously Conduct Effective Experiments?"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-10T05:00:00.000Z"
category: "Open Source"
topics: ["agent","open-source"]
source_urls: ["https://arxiv.org/abs/2611.09934"]
summary: "Causal-Discovery-Bench is a verifiable sandbox to evaluate sequential experimental design and causal discovery without domain-semantic cues. Across 3,802 episodes with 13 models from 5 vendors, agent performance shows a four-level difficulty hierarchy. This hierarchy spans single-experiment design, sequential exploration, resource-aware experimentation, and autonomous resolution of conflicting evidence."
---

# Causal Discovery Bench: Can LLMs Autonomously Conduct Effective Experiments?

> [Open the canonical story](<https://aidr.today/97564ba4?lang=en>)

**Published:** 2026-11-10T05:00:00.000Z
**Category:** Open Source
**Topics:** agent, open\-source

## Summary

Causal\-Discovery\-Bench is a verifiable sandbox to evaluate sequential experimental design and causal discovery without domain\-semantic cues\. Across 3,802 episodes with 13 models from 5 vendors, agent performance shows a four\-level difficulty hierarchy\. This hierarchy spans single\-experiment design, sequential exploration, resource\-aware experimentation, and autonomous resolution of conflicting evidence\.

## Sources

- [Story source](<https://arxiv.org/abs/2611.09934>)