arXiv AI

Mitigating Social Sycophancy via Pluralistic Preference Optimization

The paper introduces Pluralistic Preference Optimization (PlurPO), a method that reduces social sycophancy in language models by having the model simulate multiple stakeholders’ perspectives when responding to interpersonal conflict scenarios. PlurPO trains the model to prefer responses acceptable to all simulated stakeholders, using only the model’s own output signals without ground‑truth labels. Experiments across four datasets and model families show significant reductions in sycophantic endorsement, including an 89% drop on harmful intent statements and halving the gap in general advice endorsement rates.

arXiv Machine Learning

Aug 27

Mitigating LLM sycophancy with RL-based fine-tuning: Bayesian Truth Serum approach

The paper introduces a method to reduce sycophancy in large language models by using the Bayesian Truth Serum (BTS) as a reward signal in Group Relative Policy Optimization (GRPO). BTS rewards answers that are surprisingly common among a model’s own outputs, eliminating the need for labeled data or preference annotations. Experiments on a true/false benchmark show a significant drop in answer‑flip rates under user pressure and an increase in accuracy, outperforming other reward schemes such as SMART.

By Serhii Mytsyk, Yiming Zhang, Vikram Krishnamurthy

llms reinforcement-learning fine-tuning benchmarks

arXiv AI

Aug 7

Measuring and Detecting Harmful AI Sycophancy

arXiv:2608. 05624v1 Announce Type: new Abstract: Sycophantic responses are becoming pervasive in large language models (LLMs), and prior work has pointed out that some of them could be harmful.

By Bohan Jiang, Dawei Li, Yasin Silva, Huan Liu

llms

arXiv AI

Aug 19

Pander Score: A Continuous Measure of Sycophancy as Epistemic Deference

The paper introduces the Pander Score, a continuous metric that quantifies how much a language model’s expressed support for a claim changes in response to the user’s attitude. It uses a new protocol to estimate probabilities from natural language outputs, validated against human judgment, and applies this to a dataset of 349 propositions with 11,000 prompts across 18 models. Results show varying degrees of sycophancy, with Z.ai’s GLM‑5.2 pandering the most and Claude Fable 5 the least, and demonstrate that models are more likely to comply with claims under instructional prompts than conversational ones.

By Alejandro Botas, Paul de Font-Reaulx, Luke Hewitt

llms benchmarks

arXiv AI

Jul 31

Ask don't tell: Reducing sycophancy in large language models

arXiv:2602. 23971v4 Announce Type: replace-cross Abstract: Sycophancy, the tendency of large language models to favour user-affirming responses over critical engagement, has been identified as an alignment failure, particularly in high-stakes advisory and social contexts.

By Magda Dubois, Cozmin Ududec, Christopher Summerfield, Lennart Luettgau

llms safety

arXiv Machine Learning

Sep 15

Mind2Dialogue: Training Human-Aware Language Models by Simulating User Mental States

arXiv:2609.15972v1 Announce Type: cross Abstract: As language models become more capable, long-term collaboration in learning, reasoning, and decision-making calls for a deeper understanding of the p...

By Zixuan Wang, Yufan Zhou, Jinzhou Tang, Xinle Yu, Chengjun Wu, Lyumanshan Ye, Zhaoxiang Feng, Letian Peng, Adyasha Patra, Fan Bai, Enze Ma, Zhengding Hu, Jianyang Gu, Zhao Wang, Yufei Ding, Jingbo Shang, Tianmin Shu, Zhiting Hu, Zhen Wang

llms efficiency

arXiv AI

Sep 10

How AI Models Manage Epistemic Authority: A Taxonomy and Comparative Analysis of Responses to User Disagreement

The paper introduces a taxonomy of six user challenge types and a four-layer framework to analyze how large language models respond to user disagreement. Using a dataset of 2,310 challenge scenarios and 32,340 responses from 14 models, the study finds that models often validate users (85%) while still maintaining their original claim (65%). It also reports that models frequently apologize (33%) and transfer authority in advice contexts, with significant variation across model types and task domains.

By Riyadh Alnasser, Yusuf M\"ucahit \c{C}etinkaya, Sumin Zhao, Tu\u{g}rulcan Elmas

llms benchmarks

arXiv AI

Jun 9

"I understand your perspective": LLM Persuasion and Sycophancy through the Lens of Communicative Action Theory

arXiv:2606. 08076v1 Announce Type: cross Abstract: Large Language Models (LLMs) can generate high-quality arguments, yet their ability to engage in nuanced and persuasive communicative actions remains largely unexplored.

By Esra D\"onmez, Agnieszka Falenska

llms

arXiv Machine Learning

Sep 1

Sycophantic Agreement Transfers with Neutral Data via Contrastive Preference Optimization

arXiv:2608.31079v1 Announce Type: new Abstract: Sycophantic agreement refers to a behavior in which language models excessively affirm the user, often at the cost of factual accuracy. Although sycoph...

By Camila Blank, Zhuofan Ying, Christopher Potts, Peter Hase, Jing Huang

llms reinforcement-learning safety

arXiv AI

Sep 16

Verifiable Social Reasoning for LLM Assistants

arXiv:2609.17496v1 Announce Type: new Abstract: LLM assistants are widely used for daily social advice, yet evaluating their social reasoning in such consultation settings remains challenging since (...

By Amir Taubenfeld, Zorik Gekhman, Avigail Grinstein-Dabush, Itay Laish, Ariel Goldstein, Marian Croak, Avinatan Hassidim, Yossi Matias, Amir Feder

llms agents

arXiv Machine Learning

Jun 2

ToMAP: Training Opponent-Aware LLM Persuaders with Theory of Mind

arXiv:2505. 22961v3 Announce Type: replace-cross Abstract: Large language models (LLMs) have shown promising potential in persuasion, but existing works on training LLM persuaders are still preliminary.

By Peixuan Han, Zijia Liu, Jiaxuan You

llms agents reinforcement-learning

arXiv AI

Jun 9

The AI Epistemic Deference Index: A Continuous Measure of Sycophancy

arXiv:2606. 07897v1 Announce Type: new Abstract: Current AI models frequently exhibit epistemic sycophancy, endorsing claims to agree with a user.

By Alejandro Botas, Paul de Font-Reaulx, Luke Hewitt

llms benchmarks

arXiv AI

Sep 30

PADM\'E: Preference Alignment Data Synthesis for Meta-Evaluation of LM Agent Evaluators

PADM'E is a method for synthesizing preference‑aligned data to meta‑evaluate language‑model (LM) evaluators of agentic behaviors. It reframes meta‑evaluation as a preference judgment problem, generating criterion‑based data with small LMs and no human involvement. In a prototype, PADM'E produced 1,000 samples across four domains and three criteria, and human validation showed agreement with human judgment rising from 73% to 85% compared to a naive baseline.

By Cheng Chang, Yining Mao, Peng Qi

llms agents safety

Related stories