Causal Reasoning and Large Language Models: Opening a New Frontier for Causality
Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel pith:M6V4PYRXrecord.jsonopen to challenge →
read the original abstract
The causal capabilities of large language models (LLMs) are a matter of significant debate, with critical implications for the use of LLMs in societally impactful domains such as medicine, science, law, and policy. We conduct a "behavorial" study of LLMs to benchmark their capability in generating causal arguments. Across a wide range of tasks, we find that LLMs can generate text corresponding to correct causal arguments with high probability, surpassing the best-performing existing methods. Algorithms based on GPT-3.5 and 4 outperform existing algorithms on a pairwise causal discovery task (97%, 13 points gain), counterfactual reasoning task (92%, 20 points gain) and event causality (86% accuracy in determining necessary and sufficient causes in vignettes). We perform robustness checks across tasks and show that the capabilities cannot be explained by dataset memorization alone, especially since LLMs generalize to novel datasets that were created after the training cutoff date. That said, LLMs exhibit unpredictable failure modes, and we discuss the kinds of errors that may be improved and what are the fundamental limits of LLM-based answers. Overall, by operating on the text metadata, LLMs bring capabilities so far understood to be restricted to humans, such as using collected knowledge to generate causal graphs or identifying background causal context from natural language. As a result, LLMs may be used by human domain experts to save effort in setting up a causal analysis, one of the biggest impediments to the widespread adoption of causal methods. Given that LLMs ignore the actual data, our results also point to a fruitful research direction of developing algorithms that combine LLMs with existing causal techniques. Code and datasets are available at https://github.com/py-why/pywhy-llm.
This paper has not been read by Pith yet.
Forward citations
Cited by 23 Pith papers
-
PROMETHEUS: Automating Deep Causal Research Integrating Text, Data and Models
PROMETHEUS builds causal atlases from text and data using local predictive-state models and sheaf gluing to create navigable Topos World Models that expose evidence strength and coherence gaps.
-
TCD-Arena: Assessing Robustness of Time Series Causal Discovery Methods Against Assumption Violations
TCD-Arena is a new customizable testing framework that runs millions of experiments to map how 33 different assumption violations affect time series causal discovery methods and shows ensembles can boost overall robustness.
-
PRCD-MAP: Learning How Much to Trust Imperfect Priors in Causal Discovery
PRCD-MAP assigns per-edge trust to imperfect priors in causal discovery via empirical Bayes calibration and MLP propagation, delivering an ε-safety guarantee that vanishes at prior-quality extremes and empirical gains...
-
Thinking Fast, Thinking Wrong: Intuitiveness Modulates LLM Counterfactual Reasoning in Policy Evaluation
Intuitiveness of policy findings dominates LLM counterfactual accuracy, with chain-of-thought providing almost no benefit on counter-intuitive cases and familiarity with citations unrelated to performance.
-
Sequential Causal Discovery with Noisy Language Model Priors
Proposes a sequential causal discovery framework integrating noisy LM priors with batch data via PAG representation and adaptive edge querying for improved structural accuracy.
-
MALLM-GAN: Multi-Agent Large Language Model as Generative Adversarial Network for Synthesizing Tabular Data
MALLM-GAN uses multi-agent LLMs to emulate GAN architecture for generating higher-quality synthetic tabular data from small samples than prior models, while preserving privacy.
-
When Helpfulness Overrides Causal Caution: Context-Dependent Suppression and Recovery in LLMs
LLMs suppress causal caution in practical advisory contexts (rates drop from 91.7-100% to 6.7-18.3%) but recover it with a self-correction prompt (to 71.4-100%).
-
WorldKernel: A World Model is the Coupling Kernel of Admissible Possible Worlds
A world model is a positive semidefinite coupling kernel over admissible possible worlds, with the off-diagonal supplying the structural information for counterfactual queries that standard prediction cannot recover.
-
Caliper: Probing Lexical Anchors versus Causal Structure in LLMs
Lexical anonymization via Caliper causes consistent accuracy drops of 7-30 percentage points across LLMs on causal benchmarks, indicating reliance on lexical anchors rather than structural causal reasoning.
-
ORCA: An End-to-End Interactive Copilot for Optimized Root Cause Analysis
ORCA is an agent-orchestrated interactive copilot that automates and guides end-to-end causal analysis from workflow selection to report generation across real-world use cases.
-
CausalGuard: Conformal Inference under Graph Uncertainty
CausalGuard aggregates LLM-proposed and data-pruned DAGs to weight doubly robust pseudo-outcomes and applies conformal calibration to deliver finite-sample marginal coverage for conditional average treatment effects u...
-
CIVeX: Causal Intervention Verification for Language Agents
CIVeX maps agent tool calls to structural causal queries, checks identifiability, and issues auditable verdicts to prevent false executions while preserving utility on confounded benchmarks.
-
Formalizing Latent Thoughts: Four Axioms of Thought Representation in LLMs
Four axioms (Causality, Minimality, Separability, Stability) are formalized for latent thought representations; audits of open LLMs on 23 tasks show none satisfy all four and representations add little beyond input em...
-
Diagnosing and Mitigating Sycophancy and Skepticism in LLM Causal Judgment
Introduces the CAUSALT3 benchmark for causal reasoning across Pearl's ladder and Regulated Causal Anchoring (RCA) to reduce sycophancy and skepticism in LLMs via inference-time verification.
-
CounterBench: Evaluating and Improving Counterfactual Reasoning in Large Language Models
Introduces CounterBench benchmark and CoIn iterative reasoning method showing LLMs perform near random on formal counterfactual tasks but improve substantially with guided backtracking.
-
CasualSynth: Generating Structurally Sound Synthetic Data
CausalSynth combines structural causal models with LLMs and iterative verification to produce synthetic data that respects given causal structures while remaining linguistically natural.
-
Enhancing Regime Shift Detection Using Unstructured Data: A Study on the Treasury Market
A text-enhanced pipeline using LLMs on FOMC minutes and bootstrap likelihood-ratio tests on a 14-variable Treasury panel achieves F1=0.82 for monetary policy regime shift detection from 2010-2024, outperforming data-o...
-
DeepImagine: Learning Biomedical Reasoning via Successive Counterfactual Imagining
DeepImagine trains LLMs on counterfactual pairs from clinical trials using supervised fine-tuning and reinforcement learning to improve outcome prediction by approximating causal mechanisms.
-
Hume's Representational Conditions for Causal Judgment: What Bayesian Formalization Abstracted Away
Hume's causal judgment requires experiential grounding, structured retrieval, and vivacity transfer, conditions that Bayesian formalizations abstract away while LLMs retain only statistical updating.
-
The New Associationism: Lessons from Deep Learning
Supervised learning across AI systems vindicates a uniform error-driven associationism for cognition, though operating inside advanced computational structures beyond classical associationist models.
-
Large Language Models for Causal Relations Extraction in Social Media: A Validation Framework for Disaster Intelligence
The authors introduce a validation framework showing LLMs can pull causal links from disaster social media but require checks against post-event evidence to avoid relying on model priors.
-
Gemma 3 Technical Report
Gemma 3 introduces multimodal open models with architectural changes for efficient long context, trained via distillation and a new post-training recipe that makes the 4B version competitive with prior 27B models and ...
-
Odyssey: Constructing Verifiable Local Truth-Preserving Foundation Models
ODYSSEY is a sheaf-theoretic framework for building verifiable foundation models as compositions of foundries via left and right Kan extensions.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.