mlsanityData to 14 Oct 2026

Counterfactual reasoning: Testing language models' understanding of hypothetical scenarios

arXiv 2305.16572 ↗First posted 26 May 2023cs.CLcited by 27

On the map: LLM reasoning → Causal reasoning in LLMs

Abstract (arXiv)

Current pre-trained language models have enabled remarkable improvements in downstream tasks, but it remains difficult to distinguish effects of statistical correlation from more systematic logical reasoning grounded on the understanding of real world. We tease these factors apart by leveraging counterfactual conditionals, which force language models to predict unusual consequences based on hypothetical propositions. We introduce a set of tests from psycholinguistic experiments, as well as larger-scale controlled datasets, to probe counterfactual predictions from five pre-trained language models. We find that models are consistently able to override real-world knowledge in counterfactual scenarios, and that this effect is more robust in case of stronger baseline world knowledge -- however, we also find that for most models this effect appears largely to be driven by simple lexical cues. When we mitigate effects of both world knowledge and lexical cues to test knowledge of linguistic nuances of counterfactuals, we find that only GPT-3 shows sensitivity to these nuances, though this sensitivity is also non-trivially impacted by lexical associative factors.

Key facts. Cited by 27 papers in the map (data through 14 Oct 2026). Placed in Causal reasoning in LLMs, one of 17 directions in LLM reasoning. Builds on 10 of its 21 references that are in the map. First posted on arXiv 26 May 2023.

Builds on

10 of 21 references.

References · 10

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020Language Models are Few-Shot Learnerscited by 32,419
  3. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  4. 2019HellaSwag: Can a Machine Really Finish Your Sentence?cited by 4,100
  5. 2020MPNet: Masked and Permuted Pre-training for Language Understandingcited by 981
  6. 2019Explain Yourself! Leveraging Language Models for Commonsense Reasoningcited by 367
  7. 2022Language Models Are Greedy Reasoners: A Systematic Formal Analysis of Chain-of-Thoughtcited by 337
  8. 2019Probing Neural Network Comprehension of Natural Language Argumentscited by 214
  9. 2019Counterfactual Story Reasoning and Generationcited by 88
  10. 2021CRASS: A Novel Data Set and Benchmark to Test Counterfactual Reasoning of Large Language Modelscited by 34

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 5; 2024: 9; 2025: 7; 2026: 6.

27 papers.

Sort:

By area:

  1. 2023Reasoning or Reciting? Exploring the Capabilities and Limitations of Language Models Through Counterfactual Taskscited by 246
  2. 2024Beyond Accuracy: Evaluating the Reasoning Behavior of Large Language Models -- A Surveycited by 83
  3. 2023Characterizing Mechanisms for Factual Recall in Language Modelscited by 48
  4. 2025Unveiling Causal Reasoning in Large Language Models: Reality or Mirage?cited by 38
  5. 2023Examining LLMs' Uncertainty Expression Towards Questions Outside Parametric Knowledgecited by 32
  6. 2024Controllable Context Sensitivity and the Knob Behind Itcited by 23
  7. 2023MenatQA: A New Dataset for Testing the Temporal Comprehension and Reasoning Abilities of Large Language Modelscited by 15
  8. 2024A Theory for Length Generalization in Learning to Reasoncited by 11
  9. 2024MARS: Benchmarking the Metaphysical Reasoning Abilities of Language Models with a Multi-task Evaluation Datasetcited by 11
  10. 2023Conditions for Length Generalization in Learning Reasoning Skillscited by 7
  11. 2024Pre-Trained Language Models Represent Some Geographic Populations Better Than Otherscited by 6
  12. 2024The Odyssey of Commonsense Causality: From Foundational Benchmarks to Cutting-Edge Reasoningcited by 5
15 more
  1. 2025On the Eligibility of LLMs for Counterfactual Reasoning: A Decompositional Studycited by 5
  2. 2025Counterfactual reasoning: an analysis of in-context emergencecited by 4
  3. 2025GraphLocator: Graph-guided Causal Reasoning for Issue Localizationcited by 4
  4. 2024The Base-Rate Effect on LLM Benchmark Performance: Disambiguating Test-Taking Strategies from Benchmark Performancecited by 3
  5. 2025Exploring the Hidden Reasoning Process of Large Language Models by Misleading Themcited by 3
  6. 2026CaSKG: Counterfactual-Causal Skill Graphs for Scalable Agent Skill Retrievalcited by 3
  7. 2026Evidence-Grounded Verified Agentic Reasoning: A Path Toward Eliminating LLM Hallucination in Empirical Inference via Tool-Attested Kernel Proofscited by 2
  8. 2024Counterfactual Reasoning with Knowledge Graph Embeddingscited by 1
  9. 2024ACCORD: Closing the Commonsense Measurability Gapcited by 1
  10. 2025Robustness of Neurosymbolic Reasoners on First-Order Logic Problemscited by 1
  11. 2025Catch Me If You Can: How Smaller Reasoning Models Pretend to Reason with Mathematical Fidelitycited by 0
  12. 2026Evaluating LLMs as Human Surrogates in Controlled Experimentscited by 0
  13. 2026CauSim: Scaling Causal Reasoning with Increasingly Complex Causal Simulatorscited by 0
  14. 2026One mechanism for many mental spaces: a shared router over a value slot in language modelscited by 0
  15. 2026Can AI Scientists Change Their Minds? Prior-Evidence Conflict in Synthetic Universescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2022Counterfactual reasoning: Do language models need world knowledge for causal understanding?
  2. 2024Evidence from counterfactual tasks supports emergent analogical reasoning in large language models
  3. 2023What If the TV Was Off? Examining Counterfactual Reasoning Abilities of Multi-modal Language Models
  4. 2024Using Counterfactual Tasks to Evaluate the Generality of Analogical Reasoning in Large Language Models
  5. 2025Flying Pigs, FaR and Beyond: Evaluating LLM Reasoning in Counterfactual Worlds
  6. 2025CounterBench: Evaluating and Improving Counterfactual Reasoning in Large Language Models
  7. 2023Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations
  8. 2023CLOMO: Counterfactual Logical Modification with Large Language Models