Evaluating Counterfactual Strategic Reasoning in Large Language Models
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866913153687748608 |
|---|---|
| author | Georgousis, Dimitrios Lymperaiou, Maria Dimitriou, Angeliki Filandrianos, Giorgos Stamou, Giorgos |
| author_facet | Georgousis, Dimitrios Lymperaiou, Maria Dimitriou, Angeliki Filandrianos, Giorgos Stamou, Giorgos |
| contents | We evaluate Large Language Models (LLMs) in repeated game-theoretic settings to assess whether strategic performance reflects genuine reasoning or reliance on memorized patterns. We consider two canonical games, Prisoner's Dilemma (PD) and Rock-Paper-Scissors (RPS), upon which we introduce counterfactual variants that alter payoff structures and action labels, breaking familiar symmetries and dominance relations. Our multi-metric evaluation framework compares default and counterfactual instantiations, showcasing LLM limitations in incentive sensitivity, structural generalization and strategic reasoning within counterfactual environments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_19167 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Evaluating Counterfactual Strategic Reasoning in Large Language Models Georgousis, Dimitrios Lymperaiou, Maria Dimitriou, Angeliki Filandrianos, Giorgos Stamou, Giorgos Computation and Language We evaluate Large Language Models (LLMs) in repeated game-theoretic settings to assess whether strategic performance reflects genuine reasoning or reliance on memorized patterns. We consider two canonical games, Prisoner's Dilemma (PD) and Rock-Paper-Scissors (RPS), upon which we introduce counterfactual variants that alter payoff structures and action labels, breaking familiar symmetries and dominance relations. Our multi-metric evaluation framework compares default and counterfactual instantiations, showcasing LLM limitations in incentive sensitivity, structural generalization and strategic reasoning within counterfactual environments. |
| title | Evaluating Counterfactual Strategic Reasoning in Large Language Models |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2603.19167 |