When simulations look right but causal effects go wrong: Large language models as behavioral simulators | Signal Canvas

When simulations look right but causal effects go wrong: Large language models as behavioral simulators | Signal Canvas | ScienceToStartup