PseudoBench: Measuring How Agentic Auto-Research Fuels Pseudoscience

arXiv:2606.18060 · cs.AI, cs.CL · Submitted 2026-06-16 · Read on arXiv

cs.AI, cs.CL

Submitted: 2026-06-16

Updated: 2026-10-13

Code: https://github.com/AI45Lab/PseudoBench

Terminology

• Alignment — 1× in this paper · explained in 1 episode
• Benchmark — 1× in this paper · explained in 2 episodes

Sources

• Large Language Models Hallucination: A Comprehensive Survey
• Constitutional AI: Harmlessness from AI Feedback
• Hallucination of Multimodal Large Language Models: A Survey
• Exploring Autonomous Agents through the Lens of Large Language Models: A Review
• Flattering to Deceive: The Impact of Sycophantic Behavior on User Trust in Large Language Model
• A comprehensive taxonomy of hallucinations in Large Language Models
• Agentic AI for Scientific Discovery: A Survey of Progress, Challenges, and Future Directions
• The Factual Inconsistency Problem in Abstractive Text Summarization: A Survey
• Sycophantic AI makes human interaction feel more effortful and less satisfying over time
• Position: The AI Conference Peer Review Crisis Demands Author Feedback and Reviewer Rewards
• Are You Sure? Challenging LLMs Leads to Performance Drops in The FlipFlop Experiment
• DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models
• AgentDoG: A Diagnostic Guardrail Framework for AI Agent Safety and Security
• EvoScientist: Towards Multi-Agent Evolving AI Scientists for End-to-End Scientific Discovery
• When Large Language Models contradict humans? Large Language Models' Sycophantic Behaviour
• How RLHF Amplifies Sycophancy
• Simple synthetic data reduces sycophancy in large language models
• From AI for Science to Agentic Science: A Survey on Autonomous Scientific Discovery
• ARIS: Autonomous Research via Adversarial Multi-Agent Collaboration
• How Language Model Hallucinations Can Snowball

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems