Understanding the Anchoring Effect of LLM with Synthetic Data: Existence, Mechanism, and Potential Mitigations

Yiming Huang, Biquan Bie, Zuqiu NA, Weilin Ruan, Songxin Lei, Yutao Yue, Xinlei He

Keywords: Large Language Model, Anchoring Effect, Cognitive Bias, Decision-Making

Abstract:

The rise of Large Language Models (LLMs) like ChatGPT has advanced natural language processing, yet concerns about cognitive biases are growing.
In this paper, we investigate the anchoring effect, a cognitive bias where the mind relies heavily on the first information as anchors to make affected judgments.
We explore whether LLMs are affected by anchoring, the underlying mechanisms, and potential mitigation strategies.
To facilitate studies at scale on the anchoring effect, we introduce a new dataset, **_SynAnchors_** ([https://huggingface.co/datasets/TimTargaryen/SynAnchors](https://huggingface.co/datasets/TimTargaryen/SynAnchors)).
Combining refined evaluation metrics, we benchmark current widely used LLMs.
Our findings show that LLMs' anchoring bias exists commonly with shallow-layer acting and can not be eliminated by conventional strategies, while reasoning can offer some mitigation.

Paper Type: New Full Paper

Submission Number: 9