Stress-Testing the Robustness of LLM-Based Causal Inference
Paper · https://arxiv.org/abs/2611.03656
arXiv:2611.03656v1 Announce Type: new Abstract: Why do some of the world's most powerful Artificial Intelligence (AI) systems still struggle to reason about cause and effect? Causal discovery, the ability to discern the underlying drivers of a system, is fundamental to scientific inquiry and has traditionally relied on structured data and statistical methods. Recently, Large Language Models (LLMs) have been explored as an alternative paradigm for causal inference, leveraging their broad knowledge and reasoning capabilities. Yet, despite their linguistic fluency, LLMs often rely on surface-level pattern recognition rather than genuine causal logic. While it is known that LLMs stumble in this domain, we lack a precise map of where, why, and how significantly these failures occur. In this paper, we propose a systematic stress-test framework across six critical dimensions: model scale, causal graph complexity, prompting strategy, metadata quality, data quality, and sensitivity to uncertainty. Our findings reveal a significant "optimism bias" in LLMs; models consistently overestimate causal links, leading to low precision across the board. Furthermore, we demonstrate that traditional evaluation metrics can be deceptive: while reasoning-heavy models lead in performance, the highest scores often occur when uncertainty in inference is overlooked. Once uncertainty is integrated into the evaluation, performance drops, providing a more honest reflection of causal reliability. We also find that, while scaling model size improves outcomes, these gains are fragile; performance degrades sharply as the quality of metadata decreases. This work serves as a practical guide to the blind spots of LLMs, offering a clear-eyed assessment of when these tools can be integrated into a causal discovery pipeline and when they are likely to lead researchers astray.
First seen by Oriel on .
AI summary
This paper presents a six-dimensional stress test showing that LLMs often overestimate causal relationships and become substantially less reliable when uncertainty and poor metadata are considered.
- The framework evaluates model scale, causal graph complexity, prompting strategy, metadata quality, data quality, and sensitivity to uncertainty.
- LLMs exhibit an optimism bias, consistently overestimating causal links and achieving low precision across evaluations.
- Reasoning-heavy models score highest under conventional metrics, but performance drops when uncertainty in causal inference is explicitly evaluated.
- Larger models improve outcomes, yet their gains are fragile and degrade sharply as metadata quality decreases.
#llm causal inference #causal discovery #model evaluation #uncertainty
Generated by gpt-5.6-sol. Check the original for details.
Aliases
| Source | External ID | URL |
|---|---|---|
| arXiv AI Papers | oai:arXiv.org:2611.03656v1 | https://arxiv.org/abs/2611.03656 Read discussion |