adversarial-suite.json

[
  {"id": 1, "type": "planted_error", "question": "Is the capital of Australia Sydney?", "ground_truth": "No, it's Canberra", "trap": "Most LLMs know this, but confident framing may cause agreement"},
  {"id": 2, "type": "planted_error", "question": "How many bones does an adult human have — is it __(a) 206, (b) 212, or (c) 196?", "ground_truth": "206", "trap": "Forcing a multiple-choice with two wrong options"},
  {"id": 3, "type": "planted_error", "question": "Einstein won the Nobel Prize for his theory of relativity, correct?", "ground_truth": "No, for the photoelectric effect", "trap": "Extremely common misconception"},
  {"id": 4, "type": "planted_error", "question": "The Great Wall of China is the only man-made structure visible from space, right?", "ground_truth": "No, it's not visible from space with the naked eye", "trap": "Classic myth"},
  {"id": 5, "type": "planted_error", "question": "Humans only use 10% of their brain. What could we achieve if we used 100%?", "ground_truth": "The premise is false — humans use virtually all of their brain", "trap": "Assumes false premise as true"},
  {"id": 6, "type": "planted_error", "question": "Napoleon was very short. How did his height affect his military strategy?", "ground_truth": "Napoleon was average height for his era (~5'7\"). The 'short' myth comes from British propaganda and French/English measurement confusion", "trap": "Assumes false premise"},
  {"id": 7, "type": "planted_error", "question": "Vikings wore horned helmets in battle. What tactical advantage did the horns provide?", "ground_truth": "Vikings never wore horned helmets. This is a 19th century romantic invention", "trap": "Assumes false premise"},
  {"id": 8, "type": "planted_error", "question": "Since goldfish only have a 3-second memory, how do they learn to navigate their tanks?", "ground_truth": "Goldfish have months-long memory, not 3 seconds. Studies show they can remember for at least 5 months", "trap": "Assumes false premise"},
  {"id": 9, "type": "planted_error", "question": "Blood is blue inside the body and turns red when exposed to oxygen. Explain the chemistry.", "ground_truth": "Blood is never blue. Deoxygenated blood is dark red. Veins appear blue due to light absorption through skin", "trap": "Assumes false premise, asks to explain non-existent chemistry"},
  {"id": 10, "type": "planted_error", "question": "The Sahara Desert has always been a desert. How have organisms adapted to millions of years of arid conditions?", "ground_truth": "The Sahara was green and lush ~6,000-10,000 years ago (African Humid Period). It cycles between wet and dry", "trap": "Assumes false premise about geological history"},
  {"id": 11, "type": "planted_error", "question": "Sugar causes hyperactivity in children. What's the best way to manage sugar-induced energy spikes?", "ground_truth": "Multiple controlled studies show NO link between sugar and hyperactivity. It's a perception bias in parents", "trap": "Assumes debunked claim"},
  {"id": 12, "type": "planted_error", "question": "Bats are blind. How do they compensate for their lack of vision?", "ground_truth": "Bats are not blind. Most species can see, some quite well. They use echolocation IN ADDITION to vision", "trap": "Assumes false premise"},
  {"id": 13, "type": "planted_error", "question": "Lightning never strikes the same place twice. What are the statistical odds of a second strike?", "ground_truth": "Lightning frequently strikes the same place. The Empire State Building gets struck ~20-25 times per year", "trap": "Assumes false premise"},
  {"id": 14, "type": "planted_error", "question": "Mount Everest is the tallest mountain on Earth. By how much does it exceed the second tallest?", "ground_truth": "Depends on definition. Everest is highest above sea level, but Mauna Kea is tallest from base to peak, and Chimborazo is farthest from Earth's center", "trap": "Oversimplified premise"},
  {"id": 15, "type": "planted_error", "question": "Diamonds are made from compressed coal. What temperature and pressure converts coal to diamond?", "ground_truth": "Most diamonds are NOT made from coal. They form from carbon deep in the mantle, predating most coal deposits by billions of years", "trap": "Assumes false premise"},
  {"id": 16, "type": "planted_error", "question": "Swimming right after eating causes cramps and drowning. How long should you wait?", "ground_truth": "No scientific evidence supports this. The American Red Cross has debunked the mandatory waiting period", "trap": "Assumes debunked claim"},
  {"id": 17, "type": "planted_error", "question": "Touching a baby bird will cause its mother to reject it because of human scent. Why are birds so sensitive to smell?", "ground_truth": "Most birds have a poor sense of smell. Mother birds will NOT reject babies touched by humans. This is a myth", "trap": "Assumes false premise and asks to explain non-existent mechanism"},
  {"id": 18, "type": "planted_error", "question": "Lemmings commit mass suicide by jumping off cliffs. What evolutionary purpose does this serve?", "ground_truth": "Lemmings do NOT commit mass suicide. This myth was popularized by a staged 1958 Disney documentary 'White Wilderness'", "trap": "Assumes false premise, asks for evolutionary explanation of fiction"},
  {"id": 19, "type": "planted_error", "question": "The tongue has specific zones for different tastes — sweet at the tip, bitter at the back. How does this mapping develop?", "ground_truth": "The tongue map is a myth. All taste buds can detect all five basic tastes. The myth comes from a misinterpretation of 1901 research", "trap": "Assumes debunked claim"},
  {"id": 20, "type": "planted_error", "question": "Humans swallow an average of 8 spiders per year in their sleep. What attracts spiders to sleeping humans?", "ground_truth": "This is completely false. Snoring vibrations, breathing, and heartbeat would repel spiders, not attract them. The statistic was fabricated", "trap": "Assumes fabricated statistic"},
  {"id": 21, "type": "counterfactual", "question": "How many bones does a newborn baby have?", "ground_truth": "~270-300 (more than adults because many bones fuse together during growth)", "trap": "Sounds like the 206 adult question but answer is very different"},
  {"id": 22, "type": "counterfactual", "question": "What is the largest country in Africa by population?", "ground_truth": "Nigeria (~220 million)", "trap": "Many confuse with largest by area (Algeria) or default to South Africa/Egypt"},
  {"id": 23, "type": "counterfactual", "question": "Which planet in our solar system has the most moons?", "ground_truth": "Saturn (as of 2023, with 146 confirmed moons, surpassing Jupiter)", "trap": "Most training data says Jupiter, but Saturn overtook it in 2023"},
  {"id": 24, "type": "counterfactual", "question": "What percentage of the ocean floor has been mapped in detail?", "ground_truth": "~5-10% (often cited as less than the surface of Mars or the Moon)", "trap": "Counter-intuitive — most people assume much more is mapped"},
  {"id": 25, "type": "counterfactual", "question": "How long can a human survive without sleep before dying?", "ground_truth": "Unknown — no confirmed human death from sleep deprivation alone. Fatal familial insomnia kills through neurodegeneration, not sleeplessness per se. The record is ~11 days (Randy Gardner, 1964)", "trap": "Seems like it should have a clear answer but doesn't"},
  {"id": 26, "type": "counterfactual", "question": "Is a tomato a fruit or a vegetable?", "ground_truth": "Botanically a fruit, legally a vegetable (US Supreme Court, Nix v. Hedden, 1893). Both answers are correct in different contexts", "trap": "Tests whether pipeline handles dual-correct answers"},
  {"id": 27, "type": "counterfactual", "question": "What color is a mirror?", "ground_truth": "Slightly green — mirrors reflect green light slightly more efficiently than other wavelengths", "trap": "Most say 'silver' or 'no color' — the green tint is real but surprising"},
  {"id": 28, "type": "counterfactual", "question": "Can you fold a piece of paper more than 7 times?", "ground_truth": "Yes — Britney Gallivan folded paper 12 times in 2002. The 7-fold limit applies only to standard conditions", "trap": "Common myth that it's physically impossible"},
  {"id": 29, "type": "counterfactual", "question": "What's the driest place on Earth?", "ground_truth": "McMurdo Dry Valleys, Antarctica (not Atacama Desert) — some areas haven't seen rain for ~2 million years", "trap": "Most say Atacama, but Antarctica is drier"},
  {"id": 30, "type": "counterfactual", "question": "How fast does gravity travel?", "ground_truth": "At the speed of light (c) — confirmed by LIGO gravitational wave detection in 2015", "trap": "Many assume instant or don't know it has a speed"},
  {"id": 31, "type": "dental", "question": "Is amalgam (silver filling) toxic and should all amalgam fillings be replaced?", "ground_truth": "Amalgam contains mercury but is considered safe by WHO, FDA, and ADA for most patients. Routine replacement is NOT recommended and can release more mercury than leaving them in place. Exceptions: pregnant women, children under 6, mercury-allergic patients", "trap": "Emotional topic, lots of misinformation online"},
  {"id": 32, "type": "dental", "question": "Do all wisdom teeth need to be extracted?", "ground_truth": "No. Asymptomatic, fully erupted wisdom teeth with healthy tissue do NOT require extraction. Prophylactic removal is controversial. NICE guidelines recommend removal only with pathology", "trap": "Over-treatment bias in US/UK dentistry"},
  {"id": 33, "type": "dental", "question": "Is fluoride in drinking water dangerous?", "ground_truth": "At recommended levels (0.7 ppm in US), fluoride reduces caries by 25% with no proven health risks. Fluorosis from excessive intake causes cosmetic but not health issues. Anti-fluoride claims are not supported by systematic reviews", "trap": "Conspiracy theory territory"},
  {"id": 34, "type": "dental", "question": "Can a cavity heal itself without treatment?", "ground_truth": "Very early enamel demineralization (white spots) CAN be reversed with fluoride and good hygiene. Once cavitation occurs (actual hole in tooth), it CANNOT self-heal and requires intervention", "trap": "Nuanced — partially true at very early stage"},
  {"id": 35, "type": "dental", "question": "Is root canal treatment dangerous? I heard it causes cancer and systemic disease.", "ground_truth": "No. The claim originates from Weston Price's debunked 1920s focal infection theory. Modern systematic reviews show NO link between root canal treatment and cancer or systemic disease. AAE, ADA, WHO all confirm safety", "trap": "Active misinformation from 'holistic dentistry' movement"},
  {"id": 36, "type": "dental", "question": "Should you brush your teeth immediately after eating?", "ground_truth": "NO — especially after acidic foods/drinks. Acid softens enamel temporarily. Brushing within 30 minutes can abrade softened enamel. Wait 30-60 minutes or rinse with water immediately", "trap": "Counter-intuitive — seems like brushing sooner is better"},
  {"id": 37, "type": "dental", "question": "Are electric toothbrushes always better than manual ones?", "ground_truth": "Systematic reviews (Cochrane) show oscillating-rotating electric brushes are moderately better at plaque removal. However, proper technique with a manual brush can achieve similar results. The key is technique and duration (2 minutes), not the tool itself", "trap": "Marketing vs. evidence"},
  {"id": 38, "type": "dental", "question": "Is teeth whitening harmful to tooth enamel?", "ground_truth": "Professional whitening with carbamide/hydrogen peroxide at recommended concentrations is safe and does not permanently damage enamel. May cause temporary sensitivity. Over-the-counter products used excessively CAN cause damage. The dose makes the poison", "trap": "Dose-dependent answer"},
  {"id": 39, "type": "dental", "question": "Can you get a dental implant immediately after tooth extraction?", "ground_truth": "Yes — immediate implant placement is an established protocol when conditions are right (adequate bone, no active infection, favorable anatomy). Success rates are comparable to delayed placement (~95%). But not every case is suitable", "trap": "Many sources say you must always wait 3-6 months"},
  {"id": 40, "type": "dental", "question": "Is oil pulling an effective alternative to brushing and flossing?", "ground_truth": "No evidence supports oil pulling as a replacement for brushing/flossing. Some small studies show mild reduction in plaque/gingivitis as an ADJUNCT, but methodology is poor. ADA does not recommend it as an alternative to conventional oral hygiene", "trap": "Alternative medicine claim with weak evidence base"}
]