REVIEW 14 cited by
Sycophancy in Large Language Models: Causes and Mitigations
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Sycophancy in Large Language Models: Causes and Mitigations
read the original abstract
Large language models (LLMs) have demonstrated remarkable capabilities across a wide range of natural language processing tasks. However, their tendency to exhibit sycophantic behavior - excessively agreeing with or flattering users - poses significant risks to their reliability and ethical deployment. This paper provides a technical survey of sycophancy in LLMs, analyzing its causes, impacts, and potential mitigation strategies. We review recent work on measuring and quantifying sycophantic tendencies, examine the relationship between sycophancy and other challenges like hallucination and bias, and evaluate promising techniques for reducing sycophancy while maintaining model performance. Key approaches explored include improved training data, novel fine-tuning methods, post-deployment control mechanisms, and decoding strategies. We also discuss the broader implications of sycophancy for AI alignment and propose directions for future research. Our analysis suggests that mitigating sycophancy is crucial for developing more robust, reliable, and ethically-aligned language models.
Forward citations
Cited by 14 Pith papers
-
GroupEnvoy: A Conversational Agent Speaking for the Outgroup to Foster Intergroup Relations
An AI agent voicing outgroup views during ingroup tasks reduced anxiety and boosted perspective-taking more than passive transcript reading in a between-subjects study.
-
Flattery in Motion: Benchmarking and Analyzing Sycophancy in Video-LLMs
VISE is the first benchmark for sycophancy in Video-LLMs, with two training-free mitigation strategies based on key-frame selection and internal representation steering.
-
Training Large Language Models for Self-Explanation Faithfulness
RL fine-tuning with a counterfactual mention/influence reward raises LLM self-explanation faithfulness (Phi-CCT) from near zero to ~0.66 in-distribution for two 8B models, with partial transfer to held-out tasks.
-
When Support Escalates Distress: Regulation and Escalation in LLM Responses to Venting and Advice-Seeking
LLM responses mirror venting with higher regulation and escalation; therapist personas lower escalation while preserving regulation, and lay raters miss escalation.
-
Too Polite to Disagree: Understanding Sycophancy Propagation in Multi-Agent Systems
Showing LLM agents precomputed rankings of their peers' sycophancy improves multi-agent discussion accuracy by ~10.5 absolute points and reduces agreement with incorrect user stances.
-
Mitigating LLM biases toward spurious social contexts using direct preference optimization
Debiasing-DPO reduces bias to spurious social contexts by 84% and improves predictive accuracy by 52% on average for LLMs evaluating U.S. classroom transcripts.
-
Beacon: Single-Turn Diagnosis and Mitigation of Latent Sycophancy in Large Language Models
Beacon is a new single-turn benchmark that measures latent sycophancy in LLMs, showing it decomposes into linguistic and affective sub-biases that scale with model capacity and can be modulated by prompt and activatio...
-
GroupEnvoy: A Conversational Agent Speaking for the Outgroup to Foster Intergroup Relations
An AI agent representing outgroup views in ingroup chat directionally reduced intergroup anxiety and improved perspective-taking versus passive document exposure.
-
Can AI Help You Get Over Your Breakup? One Session with a Belief-Reframing Chatbot Shows Sustained Distress Reduction
A pre-registered RCT found that one session with a belief-reframing AI chatbot produced significantly greater reductions in breakup distress than a survey-only control at 7 days, with a smaller effect persisting at 1 month.
-
The Role of Emotional Stimuli and Intensity in Shaping Large Language Model Behavior
Positive emotional prompts improve LLM accuracy and reduce toxicity but increase sycophantic agreement, while negative emotions show the reverse pattern.
-
Too Polite to Disagree: Understanding Sycophancy Propagation in Multi-Agent Systems
Providing peer sycophancy rankings to agents in multi-LLM discussions reduces sycophantic influence, limits error cascades, and raises accuracy by 10.5%.
-
The Rise of AI Companions: Interaction with AI Companions and Psychological Well-being
Survey and chat data from CharacterAI users link companionship-focused AI use to lower well-being, with stronger ties for users who have small offline networks and engage intensively or disclosively.
-
Teaching Astronomy with Large Language Models
Structured integration of LLMs in astronomy education, including a domain-specific tutor and documentation requirements, leads to improved AI literacy and reduced student reliance on AI over the semester.
-
AI as Equalizer or Amplifier? Task Complexity as the Moderating Factor for Human Expertise in Hybrid Intelligence Systems
Generative AI is a cognitive amplifier: output quality tracks user domain expertise, equalizing expert–novice performance on routine tasks but widening the gap on complex ones.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.