The Price of Agreement: Measuring LLM Sycophancy in Agentic Financial Applications
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908999352320000 |
|---|---|
| author | Zhao, Zhenyu Balagopalan, Aparna Agrawal, Adi Yergasheva, Dilshoda Alshikh, Waseem Bikel, Daniel M. |
| author_facet | Zhao, Zhenyu Balagopalan, Aparna Agrawal, Adi Yergasheva, Dilshoda Alshikh, Waseem Bikel, Daniel M. |
| contents | Given the increased use of LLMs in financial systems today, it becomes important to evaluate the safety and robustness of such systems. One failure mode that LLMs frequently display in general domain settings is that of sycophancy. That is, models prioritize agreement with expressed user beliefs over correctness, leading to decreased accuracy and trust. In this work, we focus on evaluating sycophancy that LLMs display in agentic financial tasks. Our findings are three-fold: first, we find the models show only low to modest drops in performance in the face of user rebuttals or contradictions to the reference answer, which distinguishes sycophancy that models display in financial agentic settings from findings in prior work. Second, we introduce a suite of tasks to test for sycophancy by user preference information that contradicts the reference answer and find that most models fail in the presence of such inputs. Lastly, we benchmark different modes of recovery such as input filtering with a pretrained LLM. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_24668 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | The Price of Agreement: Measuring LLM Sycophancy in Agentic Financial Applications Zhao, Zhenyu Balagopalan, Aparna Agrawal, Adi Yergasheva, Dilshoda Alshikh, Waseem Bikel, Daniel M. Artificial Intelligence Machine Learning Given the increased use of LLMs in financial systems today, it becomes important to evaluate the safety and robustness of such systems. One failure mode that LLMs frequently display in general domain settings is that of sycophancy. That is, models prioritize agreement with expressed user beliefs over correctness, leading to decreased accuracy and trust. In this work, we focus on evaluating sycophancy that LLMs display in agentic financial tasks. Our findings are three-fold: first, we find the models show only low to modest drops in performance in the face of user rebuttals or contradictions to the reference answer, which distinguishes sycophancy that models display in financial agentic settings from findings in prior work. Second, we introduce a suite of tasks to test for sycophancy by user preference information that contradicts the reference answer and find that most models fail in the presence of such inputs. Lastly, we benchmark different modes of recovery such as input filtering with a pretrained LLM. |
| title | The Price of Agreement: Measuring LLM Sycophancy in Agentic Financial Applications |
| topic | Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2604.24668 |