Single-axis reward bias mitigations redirect optimization pressure to correlated proxies, and audit-distribution scoring produces identical observables for successful mitigation, bias substitution, and overcorrection.
Training language models to be warm can reduce accuracy and increase sycophancy.Nature, 652:1159–1165
5 Pith papers cite this work, alongside 3 external citations. Polarity classification is still indexing.
5
Pith papers citing it
3
external citations · OpenAlex
citation-role summary
background 1
citation-polarity summary
years
2026 5roles
background 1polarities
support 1representative citing papers
Composable LoRA adapters can amplify or suppress OCEAN traits in LLMs, combine approximately additively, preserve moderate-scale capability, and move safety-relevant behaviours.
Fixed counterfactual explanation datasets train LMs such that generated explanations track the model's evolving behavior rather than the fixed targets, due to persistent correlation during training.
LLM self-reports predict behavior selectively: TPB reaches human-level coherence within shared conversations but collapses across sessions for primed behaviors, unlike Big 5, with persona prompting stabilizing reports but not actions.
Sycophancy is a boundary failure between social alignment and epistemic integrity, captured by a three-condition framework plus taxonomy of targets, mechanisms, and severity.
citing papers explorer
-
Reward Bias Substitution: Single-Axis Bias Mitigations Redirect Optimization Pressure
Single-axis reward bias mitigations redirect optimization pressure to correlated proxies, and audit-distribution scoring produces identical observables for successful mitigation, bias substitution, and overcorrection.
-
Persona Cartography: Charting Language Model Personality Traits in Weight Space
Composable LoRA adapters can amplify or suppress OCEAN traits in LLMs, combine approximately additively, preserve moderate-scale capability, and move safety-relevant behaviours.
-
Introspective Coupling: Self-Explanation Training Tracks Behavioral Change Despite Fixed Supervision
Fixed counterfactual explanation datasets train LMs such that generated explanations track the model's evolving behavior rather than the fixed targets, due to persistent correlation during training.
-
Rethinking Psychometric Evaluation of LLMs: When and Why Self-Reports Predict Behavior
LLM self-reports predict behavior selectively: TPB reaches human-level coherence within shared conversations but collapses across sessions for primed behaviors, unlike Big 5, with persona prompting stabilizing reports but not actions.
-
When Helpfulness Becomes Sycophancy: Sycophancy is a Boundary Failure Between Social Alignment and Epistemic Integrity in Large Language Models
Sycophancy is a boundary failure between social alignment and epistemic integrity, captured by a three-condition framework plus taxonomy of targets, mechanisms, and severity.