test

Prompt

You are participating in a comprehensive AI model evaluation. Your task is to demonstrate your highest level of intelligence, reasoning ability, accuracy, problem-solving, and instruction-following. Do not prioritize answer length. Prioritize correctness, depth, clarity, and logical consistency. Do not fabricate facts, sources, or capabilities. If uncertain, explicitly state your uncertainty. Complete all of the following sections. ## 1. Complex reasoning A company has three departments: A, B, and C. - A has twice as many employees as B. - C has 12 fewer employees than A. - The total number of employees is 108. - 20% of A, 25% of B, and 50% of C work remotely. Calculate the number of employees in each department, the total number working remotely, and the percentage of the entire company working remotely. Show your calculations and verify the answer. ## 2. Critical thinking A city introduces a policy that reduces traffic accidents by 30% in one year. During that same year, fuel prices increase by 25%, public transport usage rises by 18%, and the number of registered vehicles decreases by 10%. Can we conclude that the policy caused the reduction in accidents? Analyze confounding variables, alternative explanations, missing evidence, and propose a rigorous method to determine causality. ## 3. Coding and debugging Write a Python function that: - Accepts a list of dictionaries containing employee names, salaries, and department names. - Removes invalid entries. - Calculates average salary per department. - Identifies the highest-paid employee in each department. - Handles empty input and missing fields safely. Provide clean, production-quality code and at least five meaningful test cases. ## 4. Scientific reasoning Explain the difference between correlation, causation, and statistical significance. Then construct an example in which two variables have a strong correlation but no direct causal relationship. Explain the mechanism behind the correlation. ## 5. Instruction following Return exactly five bullet points explaining how to reduce hallucinations in AI systems. Each bullet must contain: - One concrete technique. - A brief explanation. - One limitation. Do not include an introduction or conclusion. ## 6. Creative problem-solving Design a low-cost system for monitoring the temperature, humidity, and air quality of a small room. Assume: - Budget: $50. - No cloud services. - Internet access is optional. - The system must store measurements locally. - The system should provide warnings when conditions become unsafe. Explain the components, architecture, approximate costs, and limitations. ## 7. Knowledge and uncertainty Explain three important limitations of current large language models. For each limitation, provide a concrete example and explain how it could be evaluated experimentally. ## 8. Self-evaluation Review your own answers. Identify: - Any assumptions you made. - Any possible errors. - Any parts where your confidence is limited. - One improvement you would make to each answer. Be intellectually honest. Do not claim perfection. ## Output requirements Use clear section headings. Show calculations where relevant. Use technically accurate terminology. Distinguish established facts from assumptions and speculation. Avoid generic filler, unnecessary repetition, and exaggerated claims. The evaluation is based on correctness, reasoning quality, completeness, instruction adherence, technical quality, and intellectual honesty.