Reproduce “Bridging Social Psychology and LLM Reasoning: Conflict-Aware Meta-Review Generation via Cognitive Alignment” 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
Wei Chen, Han Ding, Meng Yuan, Zhao Zhang, Deqing Wang, Fuzhen Zhuang
The Cognitive Alignment Framework (CAF) addresses challenges in automated meta-review generation by integrating principles from social psychology's dual-process theory into Large Language Models. Developed by researchers from Beihang University and the Chinese Academy of Sciences, CAF achieved improvements in sentiment consistency by up to 19.47% and content consistency by up to 12.95%, alongside reducing cognitive biases like anchoring and conformity.
Key insights
- Integrating principles from dual-process theory into LLM architectures enables more nuanced, human-like reasoning, allowing LLMs to adapt their cognitive processes based on the level of conflict.
- An iterative, conflict-aware integration process is crucial for mitigating cognitive biases like the anchoring effect and conformity bias in automated meta-review generation.
- LLMs can be engineered to perform higher-order cognitive tasks like synthesizing conflicting viewpoints and deriving consensus, transforming them into adaptive arbitrators for complex scientific evaluation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.