# Unveiling Causal Reasoning in Large Language Models: Reality or Mirage?

arXiv: https://arxiv.org/abs/2506.21215 · HTML: https://mlsanity.com/paper/2506.21215/

## Key facts

- Cited by 38 papers in the map (data through 14 Oct 2026).
- Placed in Causal reasoning in LLMs, one of 17 directions in LLM reasoning.
- Builds on 48 of its 73 references that are in the map.
- First posted on arXiv 26 Jun 2025.

## Placed in

- LLM reasoning / [Causal reasoning in LLMs](https://mlsanity.com/lr/causal-reasoning-in-llms/)

## Builds on (most cited first, 30 of 48)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context](https://mlsanity.com/paper/2403.05530/) (2024)
- [Tree of Thoughts: Deliberate Problem Solving with Large Language Models](https://mlsanity.com/paper/2305.10601/) (2023)
- [Code Llama: Open Foundation Models for Code](https://mlsanity.com/paper/2308.12950/) (2023)
- [Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection](https://mlsanity.com/paper/2310.11511/) (2023)
- [Unsupervised Dense Information Retrieval with Contrastive Learning](https://mlsanity.com/paper/2112.09118/) (2021)
- [DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligence](https://mlsanity.com/paper/2401.14196/) (2024)
- [An Explanation of In-context Learning as Implicit Bayesian Inference](https://mlsanity.com/paper/2111.02080/) (2021)
- [Reasoning with Language Model is Planning with World Model](https://mlsanity.com/paper/2305.14992/) (2023)
- [LeanDojo: Theorem Proving with Retrieval-Augmented Language Models](https://mlsanity.com/paper/2306.15626/) (2023)
- [Detecting Pretraining Data from Large Language Models](https://mlsanity.com/paper/2310.16789/) (2023)
- [Causal Reasoning and Large Language Models: Opening a New Frontier for Causality](https://mlsanity.com/paper/2305.00050/) (2023)
- [Generating Wikipedia by Summarizing Long Sequences](https://mlsanity.com/paper/1801.10198/) (2018)
- [Data Distributional Properties Drive Emergent In-Context Learning in Transformers](https://mlsanity.com/paper/2205.05055/) (2022)
- [Reasoning or Reciting? Exploring the Capabilities and Limitations of Language Models Through Counterfactual Tasks](https://mlsanity.com/paper/2307.02477/) (2023)
- [MathPrompter: Mathematical Reasoning using Large Language Models](https://mlsanity.com/paper/2303.05398/) (2023)
- [Embers of Autoregression: Understanding Large Language Models Through the Problem They are Trained to Solve](https://mlsanity.com/paper/2309.13638/) (2023)
- [Language models show human-like content effects on reasoning tasks](https://mlsanity.com/paper/2207.07051/) (2022)
- [Causal Parrots: Large Language Models May Talk Causality But Are Not Causal](https://mlsanity.com/paper/2308.13067/) (2023)
- [Can Large Language Models Infer Causation from Correlation?](https://mlsanity.com/paper/2306.05836/) (2023)
- [CLadder: Assessing Causal Reasoning in Language Models](https://mlsanity.com/paper/2312.04350/) (2023)
- [Beyond Accuracy: Evaluating the Reasoning Behavior of Large Language Models -- A Survey](https://mlsanity.com/paper/2404.01869/) (2024)
- [Can large language models build causal graphs?](https://mlsanity.com/paper/2303.05279/) (2023)
- [Is ChatGPT a Good Causal Reasoner? A Comprehensive Evaluation](https://mlsanity.com/paper/2305.07375/) (2023)
- [Can LLMs Learn from Previous Mistakes? Investigating LLMs' Errors to Boost for Reasoning](https://mlsanity.com/paper/2403.20046/) (2024)

## Built on by (most cited first, 30 of 38)

- [CausalFlow: Causal Attribution and Counterfactual Repair for LLM Agent Failures](https://mlsanity.com/paper/2605.25338/) (2026)
- [Epistemological Fault Lines Between Human and Artificial Intelligence](https://mlsanity.com/paper/2512.19466/) (2025)
- [What-If World: A Causal Benchmark for General World Models in Embodied Scenarios](https://mlsanity.com/paper/2605.27589/) (2026)
- [DUAL: Learning Diverse Kernels for Aggregated Two-sample and Independence Testing](https://mlsanity.com/paper/2510.11140/) (2025)
- [Automated Penetration Testing with LLM Agents and Classical Planning](https://mlsanity.com/paper/2512.11143/) (2025)
- [Social physics in the age of artificial intelligence](https://mlsanity.com/paper/2603.16900/) (2026)
- [How to Auto-optimize Prompts for Domain Tasks? Adaptive Prompting and Reasoning through Evolutionary Domain Knowledge Adaptation](https://mlsanity.com/paper/2510.21148/) (2025)
- [Correlation or Causation: Analyzing the Causal Structures of LLM and LRM Reasoning Process](https://mlsanity.com/paper/2509.17380/) (2025)
- [YoCausal: How Far is Video Generation from World Model? A Causality Perspective](https://mlsanity.com/paper/2605.30346/) (2026)
- [Score-informed Neural Operator for Enhancing Ordering-based Causal Discovery](https://mlsanity.com/paper/2508.12650/) (2025)
- [Causally-Enhanced Reinforcement Policy Optimization](https://mlsanity.com/paper/2509.23095/) (2025)
- [Review of Hallucination Understanding in Large Language and Vision Models](https://mlsanity.com/paper/2510.00034/) (2025)
- [MemEIC: A Step Toward Continual and Compositional Knowledge Editing](https://mlsanity.com/paper/2510.25798/) (2025)
- [D2Prune: Sparsifying Large Language Models via Dual Taylor Expansion and Attention Distribution Awareness](https://mlsanity.com/paper/2601.09176/) (2026)
- [CausalT5k: Diagnosing Refusal and Failure Modes in Trustworthy Causal Reasoning Across Causal Rungs](https://mlsanity.com/paper/2602.08939/) (2026)
- [Epistemic Regret Minimization: Label-Free Causal Critique Beyond Outcome Reward](https://mlsanity.com/paper/2602.11675/) (2026)
- [Caliper: Probing Lexical Anchors versus Causal Structure in LLMs](https://mlsanity.com/paper/2606.04915/) (2026)
- [Causal Reflection with Language Models](https://mlsanity.com/paper/2508.04495/) (2025)
- [Anchor-based Maximum Discrepancy for Relative Similarity Testing](https://mlsanity.com/paper/2510.10477/) (2025)
- [Making LLMs Reliable When It Matters Most: A Five-Layer Architecture for High-Stakes Decisions](https://mlsanity.com/paper/2511.07669/) (2025)
- [Benchmarking LLMs for Pairwise Causal Discovery in Biomedical and Multi-Domain Contexts](https://mlsanity.com/paper/2601.15479/) (2026)
- [Causal Tongue-Tie: LLMs Can Encode Causal Direction, But Their Yes/No Outputs Fail to Express](https://mlsanity.com/paper/2605.25891/) (2026)
- [Why LLMs Fail at Causal Discovery and How Interventional Agents Escape](https://mlsanity.com/paper/2605.27567/) (2026)
- [SpaCellAgent: A Self-Evolving LLM-Based Multi-Agent Framework for Trajectory Analysis](https://mlsanity.com/paper/2607.07467/) (2026)
- [Leveraging Causal Reasoning Method for Explaining Medical Image Segmentation Models](https://mlsanity.com/paper/2602.20511/) (2026)
- [DAGverse: Building Document-Grounded Semantic DAGs from Scientific Papers](https://mlsanity.com/paper/2603.25293/) (2026)
- [Implicit Causal Graph Construction in Text via Chain Discovery](https://mlsanity.com/paper/2606.07525/) (2026)
- [Same Evidence, Different Target: Decoding How Diagnostic Evidence Bears on Causal Questions from Language-Model States](https://mlsanity.com/paper/2607.26929/) (2026)
- [PERO: Efficient Robust Post-Training Foundation Models for Encrypted Traffic Classification](https://mlsanity.com/paper/2608.15504/) (2026)
- [Auditing Privacy Risks in LLM-Enhanced Graph Neural Networks](https://mlsanity.com/paper/2608.25727/) (2026)

## Abstract (arXiv)

Causal reasoning capability is critical in advancing large language models (LLMs) toward strong artificial intelligence. While versatile LLMs appear to have demonstrated capabilities in understanding contextual causality and providing responses that obey the laws of causality, it remains unclear whether they perform genuine causal reasoning akin to humans. However, current evidence indicates the contrary. Specifically, LLMs are only capable of performing shallow (level-1) causal reasoning, primarily attributed to the causal knowledge embedded in their parameters, but they lack the capacity for genuine human-like (level-2) causal reasoning. To support this hypothesis, methodologically, we delve into the autoregression mechanism of transformer-based LLMs, revealing that it is not inherently causal. Empirically, we introduce a new causal Q&A benchmark called CausalProbe-2024, whose corpora are fresh and nearly unseen for the studied LLMs. The LLMs exhibit a significant performance drop on CausalProbe-2024 compared to earlier benchmarks, indicating the fact that they primarily engage in level-1 causal reasoning. To bridge the gap towards level-2 causal reasoning, we draw inspiration from the fact that human reasoning is usually facilitated by general knowledge and intended goals. We propose G^2-Reasoner, a method that incorporates general knowledge and goal-oriented prompts into LLMs' causal reasoning processes. Experiments demonstrate that G^2-Reasoner significantly enhances LLMs' causal reasoning capability, particularly in fresh and counterfactual contexts. This work sheds light on a new path for LLMs to advance towards genuine causal reasoning, going beyond level-1 and making strides towards level-2.