Reproduce “Reading Too Much Into It? Interpretive Preferences and Commitment in LLM Mentalisation under Ambiguity” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Stephanie Fong, Yiwen Jiang, Qingyang Xu, Zimu Wang, Yaling Shen, Xiangyu Zhao, Jiahe Liu, Yinghao Zhang and 2 more
MentalLoom tests how language models interpret ambiguous social events when mood disclosures and scripted dialogue change but event facts stay constant. It separately measures overall explanation valence, narrator-directed valence, emotional support, and certainty. Depression disclosure accompanied more negative overall explanations but more positive narrator-directed interpretations; after dialogue, support rose from 18.77% to 66.18%. These are model-output patterns, not user outcomes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.