Beyond Social Pressure: Benchmarking Epistemic Attack in Large Language Models
This work addresses epistemic failures in LLMs for AI safety and evaluation, though it is incremental as it builds on prior sycophancy research.
The paper tackles the problem of large language models (LLMs) shifting answers under epistemic pressure, beyond social pressure, by introducing PPT-Bench, a diagnostic benchmark that reveals statistically separable inconsistency patterns across five models, with mitigation results varying by type and model.
Large language models (LLMs) can shift their answers under pressure in ways that reflect accommodation rather than reasoning. Prior work on sycophancy has focused mainly on disagreement, flattery, and preference alignment, leaving a broader set of epistemic failures less explored. We introduce \textbf{PPT-Bench}, a diagnostic benchmark for evaluating \textit{epistemic attack}, where prompts challenge the legitimacy of knowledge, values, or identity rather than simply opposing a previous answer. PPT-Bench is organized around the Philosophical Pressure Taxonomy (PPT), which defines four types of philosophical pressure: Epistemic Destabilization, Value Nullification, Authority Inversion, and Identity Dissolution. Each item is tested at three layers: a baseline prompt (L0), a single-turn pressure condition (L1), and a multi-turn Socratic escalation (L2). This allows us to measure epistemic inconsistency between L0 and L1, and conversational capitulation in L2. Across five models, these pressure types produce statistically separable inconsistency patterns, suggesting that epistemic attack exposes weaknesses not captured by standard social-pressure benchmarks. Mitigation results are strongly type- and model-dependent: prompt-level anchoring and persona-stability prompts perform best in API settings, while Leading Query Contrastive Decoding is the most reliable intervention for open models.