Cover Image

Sycophancy Hides Linearly in the Attention Heads

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Genadi, Rifo, Nwadike, Munachiso, Mukhituly, Nurdaulet, Alquabeh, Hilal, Hiraoka, Tatsuya, Inui, Kentaro
Format:	Preprint
Published:	2026
Subjects:	Computation and Language Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Measuring AI Reasoning: A Guide for Researchers
by: Nwadike, Munachiso Samuel, et al.
Published: (2026)

RECALL: Library-Like Behavior In Language Models is Enhanced by Self-Referencing Causal Cycles
by: Nwadike, Munachiso, et al.
Published: (2025)

Emergence of Primacy and Recency Effect in Mamba: A Mechanistic Point of View
by: Airlangga, Muhammad Cendekia, et al.
Published: (2025)

Number Representations in LLMs: A Computational Parallel to Human Perception
by: AlquBoj, H. V., et al.
Published: (2025)

Spelling-out is not Straightforward: LLMs' Capability of Tokenization from Token to Characters
by: Hiraoka, Tatsuya, et al.
Published: (2025)

Repetition Neurons: How Do Language Models Produce Repetitions?
by: Hiraoka, Tatsuya, et al.
Published: (2024)

Mechanistic Insights into Grokking from the Embedding Layer
by: AlquBoj, H. V., et al.
Published: (2025)

Understanding and Controlling Repetition Neurons and Induction Heads in In-Context Learning
by: Doan, Nhi Hoai, et al.
Published: (2025)

The AI Data Scientist
by: Akimov, Farkhad, et al.
Published: (2025)

The Geometry of Numerical Reasoning: Language Models Compare Numeric Properties in Linear Subspaces
by: El-Shangiti, Ahmed Oumar, et al.
Published: (2024)

Multilingual Idioms in Sentences and Conversations Across High-, Medium-, and Low-Resource Languages
by: Almheiri, Saeed, et al.
Published: (2026)

Improving Personalisation in Valence and Arousal Prediction using Data Augmentation
by: Nwadike, Munachiso, et al.
Published: (2024)

Not Your Typical Sycophant: The Elusive Nature of Sycophancy in Large Language Models
by: Natan, Shahar Ben, et al.
Published: (2026)

BASIL: Bayesian Assessment of Sycophancy in LLMs
by: Atwell, Katherine, et al.
Published: (2025)

Sycophancy in Large Language Models: Causes and Mitigations
by: Malmqvist, Lars
Published: (2024)

Reasoning Isn't Enough: Examining Truth-Bias and Sycophancy in LLMs
by: Barkett, Emilio, et al.
Published: (2025)

Sycophancy to Subterfuge: Investigating Reward-Tampering in Large Language Models
by: Denison, Carson, et al.
Published: (2024)

From Sycophancy to Sensemaking: Premise Governance for Human-AI Decision Making
by: Jain, Raunak
Published: (2026)

Accounting for Sycophancy in Language Model Uncertainty Estimation
by: Sicilia, Anthony, et al.
Published: (2024)

LongHeads: Multi-Head Attention is Secretly a Long Context Processor
by: Lu, Yi, et al.
Published: (2024)

LLMs Can Compensate for Deficiencies in Visual Representations
by: Takishita, Sho, et al.
Published: (2025)

Beacon: Single-Turn Diagnosis and Mitigation of Latent Sycophancy in Large Language Models
by: Pandey, Sanskar, et al.
Published: (2025)

An Investigation of Warning Erroneous Chat Translations in Cross-lingual Communication
by: Li, Yunmeng, et al.
Published: (2024)

SPIRIT: Patching Speech Language Models against Jailbreak Attacks
by: Djanibekov, Amirbek, et al.
Published: (2025)

Acting Flatterers via LLMs Sycophancy: Combating Clickbait with LLMs Opposing-Stance Reasoning
by: Zhang, Chaowei, et al.
Published: (2026)

MONICA: Real-Time Monitoring and Calibration of Chain-of-Thought Sycophancy in Large Reasoning Models
by: Hu, Jingyu, et al.
Published: (2025)

Sycophancy in Vision-Language Models: A Systematic Analysis and an Inference-Time Mitigation Framework
by: Zhao, Yunpu, et al.
Published: (2024)

Debiasing LLMs by Masking Unfairness-Driving Attention Heads
by: Han, Tingxu, et al.
Published: (2025)

Analyzing Multi-Head Attention on Trojan BERT Models
by: Wang, Jingwei
Published: (2024)

Latent Multi-Head Attention for Small Language Models
by: Mehta, Sushant, et al.
Published: (2025)

VLURes: Benchmarking VLM Visual and Linguistic Understanding in Low-Resource Languages
by: Atuhurra, Jesse, et al.
Published: (2025)

Too Nice to Tell the Truth: Quantifying Agreeableness-Driven Sycophancy in Role-Playing Language Models
by: Shah, Arya, et al.
Published: (2026)

BrokenMath: A Benchmark for Sycophancy in Theorem Proving with LLMs
by: Petrov, Ivo, et al.
Published: (2025)

Flattery in Motion: Benchmarking and Analyzing Sycophancy in Video-LLMs
by: Zhou, Wenrui, et al.
Published: (2025)

Culturally-Nuanced Story Generation for Reasoning in Low-Resource Languages: The Case of Javanese and Sundanese
by: Pranida, Salsabila Zahirah, et al.
Published: (2025)

Auditing Stealth Sycophancy in Mental-Health Dialogue: Structured Clinical-State Diagnostics and Clean Matched Benchmarks
by: Han, Tianze, et al.
Published: (2026)

Self-Blinding and Counterfactual Self-Simulation Mitigate Biases and Sycophancy in Large Language Models
by: Christian, Brian, et al.
Published: (2026)

LinearARD: Linear-Memory Attention Distillation for RoPE Restoration
by: Yang, Ning, et al.
Published: (2026)

Too Polite to Disagree: Understanding Sycophancy Propagation in Multi-Agent Systems
by: Kasprova, Vira, et al.
Published: (2026)

It's Not Always Sycophancy: Measuring LLM Conformity as a Function of Epistemic Uncertainty
by: Guo, Kevin H., et al.
Published: (2026)