Abduction In Silico: Large Language Models Replicate Human Inference to the Best Explanation
This study investigates whether large language models (LLMs) can replicate the human abductive reasoning patterns reported in Douven and Mirabile (2018). Three experiments with 16 models from seven developers test sensitivity to explanation quality, the suppression effect of alternative explanations, and the link between quality difference and metacognitive confidence. To address concerns about da