Answers are grounded in this paper and cite their sources.
Reproducibility report evaluates decision functions in LLM-driven crowds, suggesting challenges in testing artifacts.
This report evaluates the artifacts associated with the article “Sense, Think, Act, Reflect: Distilling Fast and Interpretable Decision Functions from LLM-Driven Crowds”. The authors provided their code and instructions to reproduce the results presented in their paper on a permanent repository hosted at Zenodo. The evaluation was conducted on a Linux machine with an AMD EPYC 9754 128-Core Processor, 768 GiB RAM, and an NVIDIA RTX A5000 GPU (24GB). Since the proposed approach is very computationally expensive, the results were not reproduced, and not all artifacts were tested for functionality. Thus, the paper receives the Artifacts Available badge.
No takes yet. Share an insight, caveat, or question.
Anja Wolpers (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: