Yoshiki Takenami
Author directory2026
Mechanistically Understanding and Mitigating the Anchoring Effect in LLMs via Attention Head Intervention
Yoshiki Takenami | Yin Jou Huang | Yugo Murawaki | Koh Takeuchi | Chenhui Chu
Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
Yoshiki Takenami | Yin Jou Huang | Yugo Murawaki | Koh Takeuchi | Chenhui Chu
Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
Similarly to humans, LLMs are susceptible to the anchoring effect in price negotiations, which can lead to unfavorable or unintended outcomes. To ensure LLM safety in dialogue tasks such as negotiation, a mechanistic understanding and mitigation methods are essential. We propose four variants of a method to identify Anchoring Heads, which are the attention heads responsible for the anchoring effect, while preserving those essential for general tasks. Experimental results show that suppressing these heads mitigates the anchoring effect by over 16%, achieving complete mitigation in certain models without compromising general performance. Furthermore, the observation that suppressing random heads fails to mitigate the anchoring effect suggests that specific attention heads drive the anchoring effect by paying excessive attention to the anchor.
2025
How Does Cognitive Bias Affect Large Language Models? A Case Study on the Anchoring Effect in Price Negotiation Simulations
Yoshiki Takenami | Yin Jou Huang | Yugo Murawaki | Chenhui Chu
Findings of the Association for Computational Linguistics: EMNLP 2025
Yoshiki Takenami | Yin Jou Huang | Yugo Murawaki | Chenhui Chu
Findings of the Association for Computational Linguistics: EMNLP 2025
Cognitive biases, well studied in humans, can also be observed in LLMs, affecting their reliability in real-world applications. This paper investigates the anchoring effect in LLM-driven price negotiations. To this end, we instructed seller LLM agents to apply the anchoring effect and evaluated negotiations using not only an objective metric but also a subjective metric. Experimental results show that LLMs are influenced by the anchoring effect like humans. Additionally, we investigated the relationship between the anchoring effect and factors such as reasoning and personality. It was shown that reasoning models are less prone to the anchoring effect, suggesting that the long chain of thought mitigates the effect. However, we found no significant correlation between personality traits and susceptibility to the anchoring effect. These findings contribute to a deeper understanding of cognitive biases in LLMs and to the realization of safe and responsible application of LLMs in society.
Search