Counterfactual reasoning: Do language models need world knowledge for causal understanding?

Open in new window