eval_examples.json

[
  {
    "text": "The restaurant was okay — nothing special, but the pasta was edible and the waiter tried his best.",
    "labels": ["very positive", "positive", "neutral", "negative", "very negative"],
    "type": "multi-label",
    "expected": ["neutral"],
    "not_labels_explained": {
      "positive": "Praise is faint and hedged ('okay', 'tried his best'), not genuine enthusiasm.",
      "negative": "No complaint is made; the tone is resigned acceptance, not dissatisfaction."
    }
  },
  {
    "text": "I was hoping for more, honestly. The build quality is fine but the battery barely lasts half a day.",
    "labels": ["very positive", "positive", "neutral", "negative", "very negative"],
    "type": "multi-label",
    "expected": ["negative"],
    "not_labels_explained": {
      "neutral": "'Hoping for more' and 'barely lasts' express clear disappointment, not indifference.",
      "very negative": "The reviewer concedes 'build quality is fine', softening the overall stance."
    }
  },
  {
    "text": "This is hands down the best purchase I've ever made. Life-changing. I've already bought three more as gifts.",
    "labels": ["very positive", "positive", "neutral", "negative", "very negative"],
    "type": "multi-label",
    "expected": ["very positive", "positive"],
    "not_labels_explained": {
      "neutral": "Extreme superlatives ('best ever', 'life-changing') leave no room for indifference.",
      "negative": "Buying three more as gifts is the opposite of dissatisfaction."
    }
  },
  {
    "text": "It works. Does what it says on the box. Nothing more, nothing less.",
    "labels": ["very positive", "positive", "neutral", "negative", "very negative"],
    "type": "multi-label",
    "expected": ["neutral"],
    "not_labels_explained": {
      "positive": "No praise is offered; 'nothing more' signals absence of enthusiasm.",
      "negative": "The product meets expectations — no disappointment is expressed."
    }
  },
  {
    "text": "I want to cry. The screen cracked on day one, customer service hung up on me twice, and the refund policy is a joke.",
    "labels": ["very positive", "positive", "neutral", "negative", "very negative"],
    "type": "multi-label",
    "expected": ["very negative", "negative"],
    "not_labels_explained": {
      "neutral": "Emotional distress ('want to cry') and multiple failures leave no room for indifference.",
      "positive": "Every aspect mentioned is a complaint; nothing positive is acknowledged."
    }
  },
  {
    "text": "It's decent for the price, though I've seen better options. Wouldn't necessarily recommend it, but it's not terrible.",
    "labels": ["very positive", "positive", "neutral", "negative", "very negative"],
    "type": "multi-label",
    "expected": ["neutral"],
    "not_labels_explained": {
      "positive": "'Wouldn't necessarily recommend' and 'seen better options' undercut any positive lean.",
      "negative": "'Decent for the price' and 'not terrible' show it still meets a baseline."
    }
  },
  {
    "text": "The CEO announced record quarterly profits while simultaneously laying off 2,000 employees to cut costs.",
    "labels": ["corporate restructuring", "financial success", "employee welfare", "economic growth", "labor dispute"],
    "type": "multi-label",
    "expected": ["corporate restructuring", "financial success"],
    "not_labels_explained": {
      "employee welfare": "Layoffs are the opposite of welfare; the text describes harm, not care.",
      "economic growth": "Profits are company-specific, not macroeconomic growth.",
      "labor dispute": "No conflict or negotiation is described; the layoffs are unilateral."
    }
  },
  {
    "text": "New research shows that moderate coffee consumption may reduce the risk of Alzheimer's disease by up to 30%.",
    "labels": ["medical research", "nutrition advice", "drug development", "disease prevention", "public health policy"],
    "type": "multi-label",
    "expected": ["medical research", "disease prevention"],
    "not_labels_explained": {
      "nutrition advice": "The text reports a study finding, not a dietary recommendation.",
      "drug development": "Coffee is not a drug being developed; this is observational research.",
      "public health policy": "No policy or regulation is discussed."
    }
  },
  {
    "text": "Sure, let's just keep dumping plastic into the ocean. That'll definitely fix everything.",
    "labels": ["environmental activism", "sincere optimism", "sarcasm", "policy proposal", "scientific analysis"],
    "type": "multi-label",
    "expected": ["sarcasm", "environmental activism"],
    "not_labels_explained": {
      "sincere optimism": "'That'll definitely fix everything' is clearly ironic.",
      "policy proposal": "No concrete policy is proposed.",
      "scientific analysis": "No data or scientific reasoning is presented."
    }
  },
  {
    "text": "While the opposition raises valid concerns about cost, the long-term savings from renewable energy infrastructure far outweigh the initial investment, as demonstrated by Denmark's 40-year track record.",
    "labels": ["political argument", "scientific evidence", "emotional appeal", "balanced reporting", "policy advocacy"],
    "type": "multi-label",
    "expected": ["political argument", "policy advocacy"],
    "not_labels_explained": {
      "balanced reporting": "The author takes a clear side ('far outweigh'), this is not neutral reporting.",
      "scientific evidence": "Denmark's track record is a policy outcome, not a scientific experiment.",
      "emotional appeal": "The argument relies on data and logic, not emotion."
    }
  },
  {
    "text": "After years of training and countless sacrifices, the athlete finally stood on the Olympic podium, tears streaming down her face as the national anthem played.",
    "labels": ["athletic achievement", "personal sacrifice", "patriotism", "emotional moment", "celebrity gossip", "sports injury", "political protest", "entertainment review", "historical analysis", "travel experience"],
    "type": "multi-label",
    "expected": ["athletic achievement", "personal sacrifice", "patriotism", "emotional moment"],
    "not_labels_explained": {
      "celebrity gossip": "The text is a narrative of achievement, not tabloid speculation.",
      "sports injury": "Sacrifice here is metaphorical (time, effort), not physical injury.",
      "political protest": "The anthem scene is patriotic pride, not a protest.",
      "entertainment review": "This is a narrative account, not a critic's review.",
      "historical analysis": "The text describes a personal moment, not a historical examination.",
      "travel experience": "The Olympic venue is incidental; the focus is achievement."
    }
  },
  {
    "text": "The patient presented with persistent cough and fatigue over six weeks. CT scan revealed a 3cm mass in the right upper lobe.",
    "labels": ["clinical diagnosis", "treatment recommendation", "patient complaint", "surgical report", "lab results"],
    "type": "multi-label",
    "expected": ["clinical diagnosis"],
    "not_labels_explained": {
      "treatment recommendation": "No treatment is suggested; only symptoms and findings are reported.",
      "patient complaint": "This is a clinician's objective record, not the patient's own words.",
      "surgical report": "No surgery is described; this is diagnostic imaging.",
      "lab results": "A CT scan is imaging, not a laboratory test."
    }
  },
  {
    "text": "The Board of Directors has approved a dividend of $1.25 per share, payable on June 15 to shareholders of record as of May 31.",
    "labels": ["shareholder communication", "financial regulation", "investment advice", "corporate governance", "market analysis"],
    "type": "multi-label",
    "expected": ["shareholder communication", "corporate governance"],
    "not_labels_explained": {
      "financial regulation": "This is a corporate decision, not a regulatory action.",
      "investment advice": "Announcing a dividend is factual disclosure, not a buy/sell recommendation.",
      "market analysis": "No market trends or analysis are discussed."
    }
  },
  {
    "text": "My grandmother used to say that a watched pot never boils. I think about that every time I refresh my inbox waiting for the job offer.",
    "labels": ["wisdom", "nostalgia", "cooking advice", "job search", "family relationship", "proverb"],
    "type": "multi-label",
    "expected": ["nostalgia", "job search", "proverb"],
    "not_labels_explained": {
      "cooking advice": "The pot metaphor is figurative; no actual cooking guidance is given.",
      "wisdom": "The author is recounting a memory and applying a saying, not dispensing original wisdom.",
      "family relationship": "The grandmother is mentioned as a source of a saying, not as a relationship being discussed."
    }
  },
  {
    "text": "The defendant argued that the contract was void due to duress, but the court found no evidence of coercion and upheld the agreement.",
    "labels": ["legal ruling", "contract dispute", "criminal case", "arbitration", "legislative debate"],
    "type": "multi-label",
    "expected": ["legal ruling", "contract dispute"],
    "not_labels_explained": {
      "criminal case": "Duress in contract law is a civil matter; no criminal charges are described.",
      "arbitration": "The text describes a court ruling, not an arbitration process.",
      "legislative debate": "No legislation or lawmakers are involved."
    }
  },
  {
    "text": "We regret to inform you that your application has been unsuccessful. We encourage you to apply again in the future.",
    "labels": ["rejection letter", "formal apology", "encouragement", "complaint response", "job offer"],
    "type": "multi-label",
    "expected": ["rejection letter", "encouragement"],
    "not_labels_explained": {
      "formal apology": "'Regret to inform' is a formulaic softener, not an apology for wrongdoing.",
      "complaint response": "This responds to an application, not a complaint.",
      "job offer": "The application was unsuccessful — the opposite of an offer."
    }
  },
  {
    "text": "Bitcoin dropped 15% overnight after the SEC announced new cryptocurrency regulations targeting decentralized exchanges.",
    "labels": ["market crash", "regulatory action", "technology innovation", "investment opportunity", "fraud investigation"],
    "type": "multi-label",
    "expected": ["market crash", "regulatory action"],
    "not_labels_explained": {
      "technology innovation": "The text is about price and regulation, not technological advancement.",
      "investment opportunity": "A 15% drop is presented as a negative event, not a buying opportunity.",
      "fraud investigation": "Regulation is not the same as a fraud probe."
    }
  },
  {
    "text": "The children gathered around the fire as the old man began to speak of a time when the rivers ran with gold and the mountains touched the stars.",
    "labels": ["storytelling", "historical account", "children's education", "geological description", "travel writing"],
    "type": "multi-class",
    "expected": ["storytelling"],
    "not_labels_explained": {
      "historical account": "'Rivers ran with gold' and 'mountains touched the stars' are mythical, not historical.",
      "children's education": "The children are the audience, but this is oral narrative, not pedagogy.",
      "geological description": "Gold rivers and star-touching mountains are figurative, not geological.",
      "travel writing": "No real place is being described."
    }
  },
  {
    "text": "Mix 200g flour, 2 eggs, and 150ml milk. Whisk until smooth, then let rest for 30 minutes before cooking on a hot pan.",
    "labels": ["recipe instruction", "food review", "nutrition information", "cooking science", "shopping list"],
    "type": "multi-class",
    "expected": ["recipe instruction"],
    "not_labels_explained": {
      "food review": "No evaluation of taste or quality is given.",
      "nutrition information": "Quantities are for preparation, not nutritional data.",
      "cooking science": "No explanation of why the steps work is provided.",
      "shopping list": "A shopping list would not include method steps."
    }
  },
  {
    "text": "The teacher noticed that Mia had been unusually quiet for weeks. During recess, she sat alone drawing pictures of empty houses.",
    "labels": ["child psychology", "classroom management", "art education", "loneliness", "behavioral observation"],
    "type": "multi-label",
    "expected": ["behavioral observation", "loneliness"],
    "not_labels_explained": {
      "child psychology": "The text describes observations, not psychological analysis or intervention.",
      "classroom management": "No management strategy or discipline is discussed.",
      "art education": "Drawing is a behavioral symptom here, not an art lesson."
    }
  },
  {
    "text": "Temperatures are expected to reach 42°C by Thursday, with humidity levels above 80%. Authorities have issued a heat advisory for the metropolitan area.",
    "labels": ["weather forecast", "climate change", "public safety warning", "natural disaster", "environmental report"],
    "type": "multi-label",
    "expected": ["weather forecast", "public safety warning"],
    "not_labels_explained": {
      "climate change": "A single heat event is weather, not a climate trend analysis.",
      "natural disaster": "A heat advisory is a warning, not a disaster that has occurred.",
      "environmental report": "This is a short-term weather bulletin, not an environmental assessment."
    }
  },
  {
    "text": "I've been a loyal customer for 12 years and this is how you treat me? The new interface is unusable, support takes days to respond, and you raised prices by 40%.",
    "labels": ["customer complaint", "brand loyalty", "product review", "price comparison", "technical support request"],
    "type": "multi-label",
    "expected": ["customer complaint", "brand loyalty"],
    "not_labels_explained": {
      "product review": "This is a direct complaint to the company, not a public review for other buyers.",
      "price comparison": "The 40% increase is a grievance, not a comparison with competitors.",
      "technical support request": "The author criticizes support response time but is not requesting help."
    }
  },
  {
    "text": "The museum's new exhibit traces the evolution of jazz from its roots in New Orleans through bebop, cool jazz, and fusion to contemporary forms.",
    "labels": ["music history", "museum exhibit", "concert review", "music theory", "cultural tourism"],
    "type": "multi-label",
    "expected": ["music history", "museum exhibit"],
    "not_labels_explained": {
      "concert review": "No live performance is being evaluated.",
      "music theory": "The text traces historical evolution, not technical music concepts.",
      "cultural tourism": "The museum is a venue, but the focus is the exhibit's content, not tourism."
    }
  },
  {
    "text": "Dear Hiring Manager, I am writing to express my interest in the Senior Software Engineer position at your company. With 8 years of experience in distributed systems...",
    "labels": ["cover letter", "resignation letter", "performance review", "job posting", "professional recommendation"],
    "type": "multi-class",
    "expected": ["cover letter"],
    "not_labels_explained": {
      "resignation letter": "The author is applying, not leaving.",
      "performance review": "No evaluation of past work is being conducted.",
      "job posting": "The author is responding to a posting, not creating one.",
      "professional recommendation": "The author speaks for themselves, not on behalf of someone else."
    }
  },
  {
    "text": "The spacecraft successfully entered Mars orbit after a 7-month journey, marking only the fifth time a nation has achieved this milestone.",
    "labels": ["space exploration", "scientific discovery", "national achievement", "rocket engineering", "astronomy observation"],
    "type": "multi-label",
    "expected": ["space exploration", "national achievement"],
    "not_labels_explained": {
      "scientific discovery": "Entering orbit is an engineering achievement, not a new scientific finding.",
      "rocket engineering": "The text reports the mission outcome, not the engineering details.",
      "astronomy observation": "No astronomical observations are described."
    }
  },
  {
    "text": "He promised he'd change. He always promises. And every time, I believe him. Maybe that makes me the fool.",
    "labels": ["relationship conflict", "self-reflection", "deception", "emotional abuse", "personal diary"],
    "type": "multi-label",
    "expected": ["relationship conflict", "self-reflection"],
    "not_labels_explained": {
      "deception": "The text shows broken promises, but the author isn't accusing of intentional deceit — more of a pattern.",
      "emotional abuse": "The text conveys disappointment and self-doubt, but doesn't describe abuse.",
      "personal diary": "The introspective tone suggests diary, but it could be spoken aloud or in a letter."
    }
  },
  {
    "text": "Please ensure all fire exits are clearly marked and unobstructed. Monthly inspections are mandatory. Non-compliance will result in fines of up to $10,000.",
    "labels": ["safety regulation", "emergency procedure", "building inspection report", "legal threat", "workplace training"],
    "type": "multi-label",
    "expected": ["safety regulation"],
    "not_labels_explained": {
      "emergency procedure": "This sets ongoing compliance rules, not steps to follow during an emergency.",
      "building inspection report": "This is a regulation, not a report of findings from an inspection.",
      "legal threat": "The fine is a stated consequence of non-compliance, not a threat directed at someone.",
      "workplace training": "No training content or instructions for employees are provided."
    }
  },
  {
    "text": "The painting sold for $42 million at auction, far exceeding the estimated $15 million. The anonymous buyer is believed to be a Middle Eastern collector.",
    "labels": ["art market", "auction result", "art criticism", "investment news", "celebrity news"],
    "type": "multi-label",
    "expected": ["art market", "auction result"],
    "not_labels_explained": {
      "art criticism": "No aesthetic evaluation of the painting is offered.",
      "investment news": "The sale is reported as an art event, not as investment analysis.",
      "celebrity news": "The buyer is anonymous; no celebrity is identified."
    }
  },
  {
    "text": "WARNING: This product contains chemicals known to the State of California to cause cancer and reproductive harm.",
    "labels": ["health warning", "product recall", "legal disclaimer", "consumer advocacy", "scientific finding"],
    "type": "multi-label",
    "expected": ["health warning", "legal disclaimer"],
    "not_labels_explained": {
      "product recall": "The product is still sold; this is a mandated label, not a recall.",
      "consumer advocacy": "This is a regulatory requirement, not advocacy by a consumer group.",
      "scientific finding": "The warning references known science but is itself a legal notice."
    }
  },
  {
    "text": "The algorithm processes 10,000 transactions per second with 99.99% uptime. Latency averages 12ms at the 95th percentile.",
    "labels": ["performance benchmark", "sales pitch", "academic paper", "system documentation", "bug report"],
    "type": "multi-label",
    "expected": ["performance benchmark", "system documentation"],
    "not_labels_explained": {
      "sales pitch": "The tone is purely technical with no persuasive framing.",
      "academic paper": "No methodology, hypothesis, or citations are present.",
      "bug report": "All metrics are positive; no issue is reported."
    }
  },
  {
    "text": "I never thought I'd say this, but moving to the countryside was the best decision we ever made. The kids play outside until dark, and I actually sleep through the night now.",
    "labels": ["lifestyle change", "real estate recommendation", "parenting advice", "rural tourism", "personal testimony"],
    "type": "multi-label",
    "expected": ["lifestyle change", "personal testimony"],
    "not_labels_explained": {
      "real estate recommendation": "The author shares personal experience, not a recommendation to buy property.",
      "parenting advice": "Children are mentioned, but no advice is given to other parents.",
      "rural tourism": "This is about permanent relocation, not visiting."
    }
  },
  {
    "text": "The protein folds into a beta-sheet structure stabilized by hydrogen bonds between the backbone amide and carbonyl groups of alternating residues.",
    "labels": ["biochemistry", "medical treatment", "genetic engineering", "drug mechanism", "laboratory protocol"],
    "type": "multi-class",
    "expected": ["biochemistry"],
    "not_labels_explained": {
      "medical treatment": "Protein folding is fundamental science, not a treatment.",
      "genetic engineering": "No gene modification is described.",
      "drug mechanism": "No drug or pharmacological action is mentioned.",
      "laboratory protocol": "This describes molecular structure, not experimental steps."
    }
  },
  {
    "text": "Honestly, I don't care anymore. Do whatever you want. I'm done explaining myself.",
    "labels": ["anger", "indifference", "exhaustion", "passive aggression", "acceptance"],
    "type": "multi-label",
    "expected": ["exhaustion", "passive aggression"],
    "not_labels_explained": {
      "indifference": "'I don't care anymore' sounds indifferent but 'I'm done explaining' reveals frustration, not true apathy.",
      "anger": "The tone is resigned and dismissive, not actively hostile.",
      "acceptance": "The speaker hasn't reached peace — they're withdrawing out of frustration."
    }
  },
  {
    "text": "The committee voted 7-3 to approve the zoning change, allowing residential development on the former industrial site pending environmental review.",
    "labels": ["urban planning", "democratic process", "real estate development", "environmental regulation", "industrial policy"],
    "type": "multi-label",
    "expected": ["urban planning", "democratic process", "real estate development"],
    "not_labels_explained": {
      "environmental regulation": "The environmental review is pending — no regulation is described or enforced yet.",
      "industrial policy": "The site is being rezoned away from industrial use, not an industrial policy decision."
    }
  },
  {
    "text": "Place the patient in a supine position. Administer 2mg of midazolam IV. Wait 60 seconds before proceeding with intubation.",
    "labels": ["medical procedure", "emergency response", "nursing education", "drug prescription", "surgical report"],
    "type": "multi-class",
    "expected": ["medical procedure"],
    "not_labels_explained": {
      "emergency response": "The methodical step-by-step tone suggests a planned procedure, not an emergency.",
      "nursing education": "These are procedural steps, not educational material explaining why.",
      "drug prescription": "Midazolam administration is part of a procedure, not a prescription for a patient to fill.",
      "surgical report": "This is instructions to follow, not a report of what was done."
    }
  },
  {
    "text": "The UN Security Council condemned the attacks and called for an immediate ceasefire, though Russia and China abstained from the vote.",
    "labels": ["international diplomacy", "military conflict", "human rights violation", "peace negotiation", "geopolitical analysis"],
    "type": "multi-label",
    "expected": ["international diplomacy", "military conflict"],
    "not_labels_explained": {
      "human rights violation": "The attacks are condemned but no specific rights violations are detailed.",
      "peace negotiation": "A ceasefire call from the UN is not the same as active negotiation between parties.",
      "geopolitical analysis": "This is factual reporting of a vote, not analysis of power dynamics."
    }
  },
  {
    "text": "The startup raised $50M in Series B funding led by Andreessen Horowitz, bringing total funding to $78M. The company plans to expand into European markets by Q3.",
    "labels": ["venture capital", "business expansion", "IPO announcement", "financial reporting", "startup culture"],
    "type": "multi-label",
    "expected": ["venture capital", "business expansion"],
    "not_labels_explained": {
      "IPO announcement": "Series B is private funding, not a public offering.",
      "financial reporting": "This is a press announcement, not a financial statement or earnings report.",
      "startup culture": "The text discusses funding and strategy, not workplace culture."
    }
  },
  {
    "text": "You have the right to remain silent. Anything you say can and will be used against you in a court of law.",
    "labels": ["Miranda warning", "legal advice", "court proceeding", "police interrogation", "civil rights lecture"],
    "type": "multi-class",
    "expected": ["Miranda warning"],
    "not_labels_explained": {
      "legal advice": "This is a mandated notification, not advice from a lawyer.",
      "court proceeding": "This occurs at arrest, before any court involvement.",
      "police interrogation": "The warning precedes interrogation; it is not the interrogation itself.",
      "civil rights lecture": "This is a procedural recitation, not an educational discussion."
    }
  },
  {
    "text": "The glacier has retreated 2.3 kilometers over the past decade, exposing bedrock that hasn't seen sunlight in over 10,000 years.",
    "labels": ["climate evidence", "geological discovery", "environmental catastrophe", "polar exploration", "ice age research"],
    "type": "multi-label",
    "expected": ["climate evidence", "geological discovery"],
    "not_labels_explained": {
      "environmental catastrophe": "The retreat is presented as measured observation, not framed as catastrophe.",
      "polar exploration": "No exploration expedition is described.",
      "ice age research": "The 10,000-year bedrock is a consequence observed, not active research into ice ages."
    }
  },
  {
    "text": "I'm not saying it's aliens, but when you see lights moving in formation at 3am over a military base, you have to wonder.",
    "labels": ["conspiracy theory", "UFO sighting", "military activity", "scientific skepticism", "entertainment"],
    "type": "multi-label",
    "expected": ["UFO sighting", "conspiracy theory"],
    "not_labels_explained": {
      "military activity": "The military base is a location, but the speaker implies the lights are unexplained, not routine military ops.",
      "scientific skepticism": "The speaker is credulous ('you have to wonder'), not skeptical.",
      "entertainment": "The account presents itself as genuine, not as fiction or amusement."
    }
  },
  {
    "text": "To reset your password, click the link below within 24 hours. If you did not request this change, please contact support immediately.",
    "labels": ["account security", "phishing attempt", "customer support", "technical documentation", "privacy notice"],
    "type": "multi-label",
    "expected": ["account security"],
    "not_labels_explained": {
      "phishing attempt": "This is a standard legitimate password reset template, not a scam.",
      "customer support": "The email offers a self-service action, not direct support interaction.",
      "technical documentation": "This is a transactional email, not documentation.",
      "privacy notice": "No privacy policy or data handling information is discussed."
    }
  },
  {
    "text": "The violin concerto opens with a haunting melody in D minor, building through three movements to a triumphant finale that earned a five-minute standing ovation.",
    "labels": ["concert review", "music theory", "music history", "entertainment news", "cultural criticism"],
    "type": "multi-label",
    "expected": ["concert review"],
    "not_labels_explained": {
      "music theory": "D minor is mentioned descriptively, not analyzed theoretically.",
      "music history": "This reviews a specific performance, not the piece's historical significance.",
      "entertainment news": "A review evaluates quality; news would report the event factually.",
      "cultural criticism": "The review is appreciative, not critically examining cultural implications."
    }
  },
  {
    "text": "Every morning I wake up at 5am, meditate for 20 minutes, journal three pages, and run 5K before the kids are up. It changed my life.",
    "labels": ["morning routine", "self-help advice", "fitness program", "personal testimony", "productivity hack"],
    "type": "multi-label",
    "expected": ["morning routine", "personal testimony"],
    "not_labels_explained": {
      "self-help advice": "The author describes their own experience, not telling others what to do.",
      "fitness program": "Running is one element of a broader routine, not a fitness program.",
      "productivity hack": "The routine includes wellness activities (meditation, journaling), not productivity tricks."
    }
  },
  {
    "text": "The bridge, completed in 1937, spans 1,280 meters and was the longest suspension bridge in the world at the time of its construction.",
    "labels": ["engineering achievement", "historical fact", "travel guide", "architecture criticism", "infrastructure policy"],
    "type": "multi-label",
    "expected": ["engineering achievement", "historical fact"],
    "not_labels_explained": {
      "travel guide": "No visitor information or recommendations are provided.",
      "architecture criticism": "The text states facts without aesthetic evaluation.",
      "infrastructure policy": "No policy discussion about building or maintaining bridges."
    }
  },
  {
    "text": "Look, I get it, change is hard. But we've been doing things this way for 20 years and the market has moved on. If we don't adapt, we die. Simple as that.",
    "labels": ["business strategy", "motivational speech", "threat", "market analysis", "change management"],
    "type": "multi-label",
    "expected": ["business strategy", "change management"],
    "not_labels_explained": {
      "motivational speech": "The tone is blunt and urgent ('we die'), not inspirational or uplifting.",
      "threat": "'We die' is a metaphor for business failure, not an actual threat.",
      "market analysis": "The market is referenced but no data or trends are analyzed."
    }
  },
  {
    "text": "The accused was found not guilty on all charges. Outside the courthouse, the victim's family wept while the defense team celebrated.",
    "labels": ["court verdict", "wrongful conviction", "victim impact", "legal victory", "media coverage"],
    "type": "multi-label",
    "expected": ["court verdict", "victim impact", "legal victory"],
    "not_labels_explained": {
      "wrongful conviction": "The verdict is 'not guilty' — there is no conviction at all.",
      "media coverage": "The text describes the scene, but is not self-referentially about media."
    }
  },
  {
    "text": "Add 1 tbsp olive oil to a hot skillet. Sear the salmon skin-side down for 4 minutes. Flip and cook for 2 more minutes. Rest for 1 minute before serving.",
    "labels": ["cooking instruction", "restaurant review", "nutrition guide", "food science", "meal planning"],
    "type": "multi-class",
    "expected": ["cooking instruction"],
    "not_labels_explained": {
      "restaurant review": "No restaurant or dining experience is evaluated.",
      "nutrition guide": "No nutritional information is provided.",
      "food science": "No explanation of why the technique works.",
      "meal planning": "This is a single preparation step, not a meal plan."
    }
  },
  {
    "text": "The company's carbon emissions fell 23% year-over-year, but critics note that the reduction came primarily from selling off its most polluting subsidiaries rather than genuine operational improvements.",
    "labels": ["environmental progress", "corporate greenwashing", "emissions reporting", "investor relations", "environmental activism"],
    "type": "multi-label",
    "expected": ["corporate greenwashing", "emissions reporting"],
    "not_labels_explained": {
      "environmental progress": "The critics' point is that the reduction is accounting, not genuine progress.",
      "investor relations": "This reads as journalism, not a communication to investors.",
      "environmental activism": "The criticism comes from unnamed critics, not from activists campaigning."
    }
  },
  {
    "text": "She scored 47 points in the championship game, grabbed 12 rebounds, and dished out 8 assists — a performance many are calling the greatest in tournament history.",
    "labels": ["sports achievement", "player statistics", "game recap", "sports commentary", "talent scouting"],
    "type": "multi-label",
    "expected": ["sports achievement", "player statistics", "game recap"],
    "not_labels_explained": {
      "sports commentary": "The text reports facts and others' opinions, but isn't live or opinionated commentary.",
      "talent scouting": "No evaluation of future potential or recruitment value is discussed."
    }
  },
  {
    "text": "As a landlord, I can tell you that tenant rights have gone too far. We can't evict non-paying tenants for months, and meanwhile the mortgage still needs to be paid.",
    "labels": ["landlord perspective", "tenant rights criticism", "housing policy", "financial hardship", "legal complaint"],
    "type": "multi-label",
    "expected": ["landlord perspective", "tenant rights criticism", "financial hardship"],
    "not_labels_explained": {
      "housing policy": "The speaker criticizes policy effects but doesn't propose or analyze policy.",
      "legal complaint": "This is a personal grievance, not a formal legal filing."
    }
  },
  {
    "text": "The 3.2-magnitude earthquake was felt across the valley but caused no damage. Seismologists say it was an aftershock from last month's larger event.",
    "labels": ["seismic event", "natural disaster", "scientific explanation", "emergency alert", "geological research"],
    "type": "multi-label",
    "expected": ["seismic event", "scientific explanation"],
    "not_labels_explained": {
      "natural disaster": "A 3.2-magnitude quake with no damage doesn't qualify as a disaster.",
      "emergency alert": "No evacuation or safety instructions are issued.",
      "geological research": "Seismologists provide context, but no research is described."
    }
  },
  {
    "text": "Thank you for your years of service. Your dedication to this team has been exceptional, and while we're sad to see you go, we wish you all the best in your next chapter.",
    "labels": ["farewell message", "performance evaluation", "retirement announcement", "termination notice", "professional recommendation"],
    "type": "multi-class",
    "expected": ["farewell message"],
    "not_labels_explained": {
      "performance evaluation": "The praise is part of a goodbye, not a formal evaluation.",
      "retirement announcement": "The reason for leaving is unspecified; it may not be retirement.",
      "termination notice": "The warm tone ('sad to see you go') indicates voluntary departure.",
      "professional recommendation": "This is addressed to the person leaving, not to a third party."
    }
  },
  {
    "text": "Participants who received the placebo showed no improvement, while those on the 200mg dose reported a 40% reduction in symptoms over 12 weeks.",
    "labels": ["clinical trial results", "drug advertisement", "medical advice", "patient testimonial", "regulatory approval"],
    "type": "multi-class",
    "expected": ["clinical trial results"],
    "not_labels_explained": {
      "drug advertisement": "The tone is objective and comparative, not promotional.",
      "medical advice": "Results are reported, not prescribed as advice.",
      "patient testimonial": "Aggregate data is presented, not individual experience.",
      "regulatory approval": "Trial results inform approval but this text doesn't describe an approval decision."
    }
  },
  {
    "text": "The Wi-Fi keeps dropping every 15 minutes and the Bluetooth won't pair with any of my devices. I've tried factory reset twice. Please help.",
    "labels": ["technical support request", "product review", "bug report", "user error", "warranty claim"],
    "type": "multi-class",
    "expected": ["technical support request"],
    "not_labels_explained": {
      "product review": "'Please help' makes this a support request, not a review for other buyers.",
      "bug report": "A bug report is filed with developers; this is a user asking for help.",
      "user error": "The user has tried troubleshooting (factory reset); nothing indicates user error.",
      "warranty claim": "No mention of warranty, replacement, or return."
    }
  },
  {
    "text": "In fiscal year 2025, revenue grew 18% to $4.2B driven by cloud services, while operating margins expanded 300 basis points to 22.4%.",
    "labels": ["earnings report", "investment thesis", "market forecast", "business journalism", "accounting audit"],
    "type": "multi-class",
    "expected": ["earnings report"],
    "not_labels_explained": {
      "investment thesis": "This reports past performance, not a forward-looking investment argument.",
      "market forecast": "No predictions are made.",
      "business journalism": "The dry financial language suggests an official report, not journalism.",
      "accounting audit": "This summarizes results, not an auditor's verification process."
    }
  },
  {
    "text": "The teacher gave us exactly 45 minutes to finish the exam, but I spent 30 of those staring at question 7, which made absolutely no sense.",
    "labels": ["exam experience", "test anxiety", "teacher criticism", "time management", "academic cheating"],
    "type": "multi-label",
    "expected": ["exam experience"],
    "not_labels_explained": {
      "test anxiety": "Frustration with a confusing question is not the same as anxiety disorder.",
      "teacher criticism": "The complaint is about the question, not the teacher personally.",
      "time management": "Poor time allocation is described, but the author doesn't frame it as a time management issue.",
      "academic cheating": "Nothing suggests cheating."
    }
  },
  {
    "text": "By 2050, an estimated 2 billion people will live in areas facing severe water scarcity, according to a report published by the World Health Organization.",
    "labels": ["climate projection", "public health warning", "resource scarcity", "policy recommendation", "demographic study"],
    "type": "multi-label",
    "expected": ["climate projection", "resource scarcity", "public health warning"],
    "not_labels_explained": {
      "policy recommendation": "The report projects a problem but no policy is recommended here.",
      "demographic study": "Population is mentioned in context of water, not as a demographic analysis."
    }
  },
  {
    "text": "I know you meant well, but telling someone who just lost their job to 'look on the bright side' is really not helpful.",
    "labels": ["emotional boundary", "constructive criticism", "toxic positivity", "relationship advice", "grief counseling"],
    "type": "multi-label",
    "expected": ["emotional boundary", "toxic positivity"],
    "not_labels_explained": {
      "constructive criticism": "The speaker corrects insensitivity, but isn't offering structured feedback.",
      "relationship advice": "This is a direct personal communication, not advice for others.",
      "grief counseling": "Job loss is upsetting, but no counseling is occurring."
    }
  },
  {
    "text": "The vintage 1961 Château Latour exhibited notes of tobacco, cedar, and blackcurrant with remarkably preserved tannins despite its age.",
    "labels": ["wine tasting note", "product advertisement", "luxury lifestyle", "food pairing guide", "auction listing"],
    "type": "multi-class",
    "expected": ["wine tasting note"],
    "not_labels_explained": {
      "product advertisement": "The descriptive tone is evaluative, not promotional.",
      "luxury lifestyle": "The wine is expensive, but the text focuses on taste, not lifestyle.",
      "food pairing guide": "No food pairings are suggested.",
      "auction listing": "No price, lot number, or sale information is given."
    }
  },
  {
    "text": "The government has allocated $2 billion for rural broadband expansion, aiming to connect 10 million underserved households by 2028.",
    "labels": ["infrastructure investment", "digital inclusion", "government spending", "technology deployment", "election promise"],
    "type": "multi-label",
    "expected": ["infrastructure investment", "digital inclusion", "government spending"],
    "not_labels_explained": {
      "technology deployment": "The text announces funding, not actual deployment of technology.",
      "election promise": "Budget allocation is an enacted policy, not a campaign promise."
    }
  },
  {
    "text": "My dog ate my homework. Seriously. He literally chewed through my entire history essay. I have the shredded remains as proof.",
    "labels": ["excuse", "humor", "factual account", "academic dishonesty", "pet behavior"],
    "type": "multi-label",
    "expected": ["excuse", "factual account", "humor", "pet behavior"],
    "not_labels_explained": {
      "academic dishonesty": "The homework was destroyed by a pet, not fabricated or plagiarized."
    }
  },
  {
    "text": "The evacuation order covers all areas east of Highway 101. Residents should leave immediately and proceed to the community center on Oak Street.",
    "labels": ["emergency directive", "weather forecast", "community announcement", "travel advisory", "disaster relief"],
    "type": "multi-class",
    "expected": ["emergency directive"],
    "not_labels_explained": {
      "weather forecast": "No weather information is given; this is about immediate evacuation.",
      "community announcement": "The urgency ('leave immediately') elevates this beyond a routine announcement.",
      "travel advisory": "This orders evacuation, not advises on travel conditions.",
      "disaster relief": "Relief comes after the event; this is an active evacuation order."
    }
  },
  {
    "text": "The Renaissance saw a flourishing of art, science, and philosophy across Europe, fundamentally reshaping Western civilization's relationship with knowledge and beauty.",
    "labels": ["historical overview", "art criticism", "philosophical argument", "textbook content", "cultural analysis"],
    "type": "multi-label",
    "expected": ["historical overview", "textbook content"],
    "not_labels_explained": {
      "art criticism": "Art is mentioned as part of a broader era, not individually critiqued.",
      "philosophical argument": "No philosophical claim is argued; the text describes what happened.",
      "cultural analysis": "The overview is too broad and neutral to qualify as deep cultural analysis."
    }
  },
  {
    "text": "The union rejected management's latest offer, calling the proposed 2% raise 'insulting' given record company profits and 6% inflation.",
    "labels": ["labor negotiation", "wage dispute", "corporate greed", "economic analysis", "strike action"],
    "type": "multi-label",
    "expected": ["labor negotiation", "wage dispute"],
    "not_labels_explained": {
      "corporate greed": "'Record company profits' is context for the union's position, not a characterization by the author.",
      "economic analysis": "Inflation is cited as a bargaining point, not analyzed economically.",
      "strike action": "The offer was rejected, but no strike is mentioned."
    }
  },
  {
    "text": "Plant the seeds 1 inch deep and 6 inches apart in well-drained soil. Water daily for the first two weeks, then reduce to every other day.",
    "labels": ["gardening instruction", "botanical research", "agricultural policy", "landscape design", "nature writing"],
    "type": "multi-class",
    "expected": ["gardening instruction"],
    "not_labels_explained": {
      "botanical research": "These are practical planting steps, not research.",
      "agricultural policy": "No policy or regulation is discussed.",
      "landscape design": "Planting a specific crop is not landscape architecture.",
      "nature writing": "The imperative tone is instructional, not contemplative."
    }
  },
  {
    "text": "According to witnesses, the suspect fled on foot heading northbound on Main Street. He is described as a white male, approximately 6 feet tall, wearing a dark hoodie.",
    "labels": ["police report", "crime fiction", "witness testimony", "suspect identification", "news report"],
    "type": "multi-label",
    "expected": ["police report", "suspect identification"],
    "not_labels_explained": {
      "crime fiction": "The precise details and formal register indicate a real report.",
      "witness testimony": "Witnesses are cited, but this is the officer's summary, not direct testimony.",
      "news report": "The formal, procedural language suggests an official police document, not journalism."
    }
  },
  {
    "text": "The hotel room was clean and spacious with a beautiful ocean view. However, the walls were paper-thin and we could hear everything from the room next door.",
    "labels": ["hotel review", "travel recommendation", "complaint letter", "tourism advertising", "hospitality management"],
    "type": "multi-label",
    "expected": ["hotel review"],
    "not_labels_explained": {
      "travel recommendation": "Mixed feedback with a significant negative makes this a review, not a recommendation.",
      "complaint letter": "The balanced positive/negative tone is a review, not a formal complaint.",
      "tourism advertising": "Mentioning thin walls is not promotional.",
      "hospitality management": "The review doesn't discuss management practices."
    }
  },
  {
    "text": "I think we need to talk. It's not you, it's me. I've been feeling this way for a while now and I think it's better if we go our separate ways.",
    "labels": ["breakup conversation", "self-blame", "conflict resolution", "emotional manipulation", "couples therapy"],
    "type": "multi-label",
    "expected": ["breakup conversation", "self-blame"],
    "not_labels_explained": {
      "conflict resolution": "This ends the relationship rather than resolving conflict within it.",
      "emotional manipulation": "'It's not you, it's me' is a cliché softener, not necessarily manipulative.",
      "couples therapy": "No therapeutic context; this is a direct personal conversation."
    }
  },
  {
    "text": "The package was delivered three days late, arrived damaged, and contained the wrong item. When I called customer service, I was put on hold for 45 minutes.",
    "labels": ["customer complaint", "delivery failure", "quality control issue", "customer service review", "refund request"],
    "type": "multi-label",
    "expected": ["customer complaint", "delivery failure", "quality control issue"],
    "not_labels_explained": {
      "customer service review": "The hold time is one detail in a broader complaint, not a CS review.",
      "refund request": "The customer describes problems but doesn't explicitly request a refund."
    }
  },
  {
    "text": "The candidate has a strong track record in project management, excellent communication skills, and is well-regarded by colleagues. However, they may need additional support in stakeholder management at the executive level.",
    "labels": ["performance review", "job recommendation", "interview feedback", "promotion decision", "career coaching"],
    "type": "multi-label",
    "expected": ["performance review", "job recommendation"],
    "not_labels_explained": {
      "interview feedback": "The evaluator knows the candidate well ('track record', 'well-regarded by colleagues'), suggesting an existing relationship.",
      "promotion decision": "The text evaluates readiness but doesn't announce a promotion decision.",
      "career coaching": "The developmental note is observational, not coaching advice."
    }
  },
  {
    "text": "If you or someone you know is struggling with substance abuse, please call the National Helpline at 1-800-662-4357. Help is available 24/7, and all calls are confidential.",
    "labels": ["public service announcement", "medical advice", "addiction treatment", "crisis intervention", "government program"],
    "type": "multi-label",
    "expected": ["public service announcement", "crisis intervention"],
    "not_labels_explained": {
      "medical advice": "Directing to a helpline is not the same as giving medical advice.",
      "addiction treatment": "The PSA points to a resource, not treatment itself.",
      "government program": "The helpline may be government-funded, but the text doesn't describe a program."
    }
  },
  {
    "text": "The experiment was repeated three times under identical conditions. In all trials, the catalyst reduced reaction time by approximately 60%, confirming the initial hypothesis.",
    "labels": ["scientific finding", "experimental methodology", "peer review", "research proposal", "textbook example"],
    "type": "multi-label",
    "expected": ["scientific finding", "experimental methodology"],
    "not_labels_explained": {
      "peer review": "This reports results; peer review is an external evaluation process.",
      "research proposal": "The experiment has already been conducted, not proposed.",
      "textbook example": "This describes original research, not a pedagogical illustration."
    }
  },
  {
    "text": "This neighborhood has seen a 200% increase in property values over the past decade, driven by the new transit line and the influx of tech companies to the area.",
    "labels": ["real estate market analysis", "gentrification", "urban development", "investment advice", "community impact"],
    "type": "multi-label",
    "expected": ["real estate market analysis", "gentrification", "urban development"],
    "not_labels_explained": {
      "investment advice": "Market trends are reported, not positioned as buying advice.",
      "community impact": "The text focuses on economic metrics, not community effects on residents."
    }
  },
  {
    "text": "Ladies and gentlemen of the jury, the evidence clearly shows that on the night of March 12th, my client was 200 miles away attending his daughter's recital.",
    "labels": ["defense argument", "alibi claim", "jury instruction", "opening statement", "witness examination"],
    "type": "multi-label",
    "expected": ["defense argument", "alibi claim"],
    "not_labels_explained": {
      "jury instruction": "A lawyer is addressing the jury, but only a judge gives jury instructions.",
      "opening statement": "The definitive claim ('clearly shows') suggests closing argument, not opening.",
      "witness examination": "No witness is being questioned; the lawyer addresses the jury directly."
    }
  },
  {
    "text": "When I was eight, my father taught me to fish by the river behind our house. We never caught much, but those mornings are the happiest memories of my childhood.",
    "labels": ["nostalgia", "fishing guide", "father-child relationship", "outdoor recreation", "childhood memoir"],
    "type": "multi-label",
    "expected": ["nostalgia", "father-child relationship", "childhood memoir"],
    "not_labels_explained": {
      "fishing guide": "Fishing is the setting, not the subject; no technique is taught.",
      "outdoor recreation": "The focus is emotional memory, not the recreational activity."
    }
  },
  {
    "text": "The new tax reform will increase the standard deduction to $15,000 for individuals and $30,000 for married couples filing jointly, effective January 1.",
    "labels": ["tax policy change", "financial advice", "legislative news", "accounting update", "election issue"],
    "type": "multi-label",
    "expected": ["tax policy change", "legislative news"],
    "not_labels_explained": {
      "financial advice": "The text reports a policy change, not advice on how to file.",
      "accounting update": "This is a legal change, not a change in accounting standards.",
      "election issue": "The reform is enacted ('effective January 1'), not a campaign proposal."
    }
  },
  {
    "text": "I don't usually write reviews, but this book genuinely changed how I think about relationships. If you've ever felt stuck in unhealthy patterns, read this.",
    "labels": ["book recommendation", "self-help endorsement", "literary criticism", "personal transformation", "therapy alternative"],
    "type": "multi-label",
    "expected": ["book recommendation", "personal transformation"],
    "not_labels_explained": {
      "self-help endorsement": "The author is a reader sharing impact, not officially endorsing.",
      "literary criticism": "No analysis of writing quality, structure, or literary merit.",
      "therapy alternative": "The book changed thinking, but isn't positioned as replacing therapy."
    }
  },
  {
    "text": "Error 503: Service Unavailable. The server is temporarily unable to handle the request due to maintenance or capacity issues. Please try again later.",
    "labels": ["system error", "server maintenance", "security breach", "user error", "service outage"],
    "type": "multi-label",
    "expected": ["system error", "service outage"],
    "not_labels_explained": {
      "server maintenance": "The message says 'maintenance or capacity issues' — it could be overload, not confirmed maintenance.",
      "security breach": "503 is a capacity/maintenance issue, not a security incident.",
      "user error": "The error is server-side, not caused by the user."
    }
  },
  {
    "text": "The photograph captures a lone figure standing at the edge of a cliff, silhouetted against a blood-red sunset. The composition draws the eye downward into the void below.",
    "labels": ["art description", "landscape photography", "existential theme", "travel photography", "suicide imagery"],
    "type": "multi-label",
    "expected": ["art description", "existential theme"],
    "not_labels_explained": {
      "landscape photography": "The focus is on composition and mood, not landscape documentation.",
      "travel photography": "No location or travel context is mentioned.",
      "suicide imagery": "The void is compositional and metaphorical; no suicidal intent is described."
    }
  },
  {
    "text": "Attention all staff: the office will be closed on Friday, December 24th for the holiday. Please ensure all client deliverables are completed by Thursday end of day.",
    "labels": ["office announcement", "holiday notice", "deadline reminder", "HR policy", "event invitation"],
    "type": "multi-label",
    "expected": ["office announcement", "holiday notice", "deadline reminder"],
    "not_labels_explained": {
      "HR policy": "This is a one-time announcement, not a standing policy.",
      "event invitation": "An office closure is not an event being invited to."
    }
  },
  {
    "text": "The puppy knocked over the Christmas tree, ate half the ornaments, and then fell asleep in the wrapping paper. Merry Christmas to us.",
    "labels": ["holiday anecdote", "pet misbehavior", "home disaster", "humor", "animal cruelty"],
    "type": "multi-label",
    "expected": ["holiday anecdote", "pet misbehavior", "humor"],
    "not_labels_explained": {
      "home disaster": "The tone is affectionate and humorous, not distressed.",
      "animal cruelty": "The puppy caused chaos; it was not harmed."
    }
  },
  {
    "text": "The suspect's DNA was found on the weapon, his fingerprints were on the door handle, and cell tower data places his phone at the scene between 11pm and 2am.",
    "labels": ["forensic evidence", "prosecution argument", "crime scene analysis", "surveillance report", "confession"],
    "type": "multi-label",
    "expected": ["forensic evidence", "prosecution argument"],
    "not_labels_explained": {
      "crime scene analysis": "The evidence is compiled and presented, not being analyzed at the scene.",
      "surveillance report": "Cell tower data is one element, but this is a broader evidence summary.",
      "confession": "Physical evidence is presented; the suspect has not admitted anything."
    }
  },
  {
    "text": "Between 1990 and 2020, global extreme poverty fell from 36% to under 10%, lifting over a billion people above the $1.90/day threshold.",
    "labels": ["development progress", "economic statistic", "poverty eradication", "wealth inequality", "humanitarian aid"],
    "type": "multi-label",
    "expected": ["development progress", "economic statistic"],
    "not_labels_explained": {
      "poverty eradication": "Poverty fell dramatically but was not eradicated; 10% remain.",
      "wealth inequality": "The data shows poverty reduction, not the distribution of wealth.",
      "humanitarian aid": "The improvement may involve aid, but the text attributes no cause."
    }
  },
  {
    "text": "The council voted unanimously to ban single-use plastics in all municipal buildings, effective March 1. Reusable alternatives will be provided at no cost to employees.",
    "labels": ["environmental policy", "government regulation", "sustainability initiative", "cost-cutting measure", "employee benefit"],
    "type": "multi-label",
    "expected": ["environmental policy", "government regulation", "sustainability initiative"],
    "not_labels_explained": {
      "cost-cutting measure": "The policy provides free alternatives; cost reduction is not the motivation.",
      "employee benefit": "Free reusable items are a practical provision, not a benefit program."
    }
  },
  {
    "text": "I'm sorry, but I cannot process your request as it violates our terms of service, specifically Section 4.2 regarding prohibited content.",
    "labels": ["content moderation", "customer rejection", "censorship", "legal notice", "automated response"],
    "type": "multi-label",
    "expected": ["content moderation", "customer rejection", "automated response"],
    "not_labels_explained": {
      "censorship": "Enforcing agreed-upon terms of service is not censorship.",
      "legal notice": "This references TOS but is an operational denial, not a legal communication."
    }
  },
  {
    "text": "The choir's rendition of Handel's Messiah brought the audience to tears. The soprano's high notes were crystalline, and the bass section provided an extraordinary foundation.",
    "labels": ["music review", "concert report", "religious event", "talent assessment", "cultural commentary"],
    "type": "multi-label",
    "expected": ["music review", "concert report"],
    "not_labels_explained": {
      "religious event": "Handel's Messiah is performed as concert repertoire; no religious service is described.",
      "talent assessment": "The review praises the performance, not evaluating individual talent for recruitment.",
      "cultural commentary": "This is a specific performance review, not broader cultural commentary."
    }
  },
  {
    "text": "She walked into the interview room, sat down, and immediately asked for a lawyer. The detectives exchanged a glance and turned off the recorder.",
    "labels": ["police interrogation", "legal drama", "right to counsel", "confession scene", "detective work"],
    "type": "multi-label",
    "expected": ["police interrogation", "right to counsel"],
    "not_labels_explained": {
      "legal drama": "The factual register suggests a real account, not fiction.",
      "confession scene": "The opposite occurs — she refuses to speak without a lawyer.",
      "detective work": "The detectives stop; no investigation technique is demonstrated."
    }
  },
  {
    "text": "Remember when we used to ride bikes to the corner store for penny candy? Those days are gone forever, and something precious went with them.",
    "labels": ["nostalgia", "social commentary", "childhood memory", "cultural decline", "personal loss"],
    "type": "multi-label",
    "expected": ["nostalgia", "childhood memory"],
    "not_labels_explained": {
      "social commentary": "The lament is personal and emotional, not a sociological observation.",
      "cultural decline": "The speaker mourns personal loss of innocence, not cultural deterioration.",
      "personal loss": "Nothing concrete was lost; the speaker mourns the passage of time."
    }
  },
  {
    "text": "According to our records, your subscription will expire on June 30. To continue uninterrupted service, please update your payment method by June 25.",
    "labels": ["subscription renewal", "billing notice", "service termination", "promotional offer", "account warning"],
    "type": "multi-label",
    "expected": ["subscription renewal", "billing notice"],
    "not_labels_explained": {
      "service termination": "The notice warns of expiration and offers renewal, not termination.",
      "promotional offer": "No discount or special offer is mentioned.",
      "account warning": "This is a routine billing notice, not a warning about account misuse."
    }
  },
  {
    "text": "The volcano last erupted in 1783, killing over 9,000 people and causing a famine that lasted two years. Scientists say another eruption is overdue.",
    "labels": ["volcanic history", "disaster warning", "scientific prediction", "geological study", "emergency preparedness"],
    "type": "multi-label",
    "expected": ["volcanic history", "scientific prediction"],
    "not_labels_explained": {
      "disaster warning": "Scientists note it's 'overdue', but no official warning is issued.",
      "geological study": "Historical facts and a prediction are mentioned, not a study's methodology.",
      "emergency preparedness": "No preparedness measures or plans are discussed."
    }
  },
  {
    "text": "Preheat oven to 180°C. In a large bowl, cream together butter and sugar until light and fluffy. Add eggs one at a time, beating well after each addition.",
    "labels": ["baking recipe", "cooking chemistry", "food blog", "kitchen equipment review", "nutrition information"],
    "type": "multi-class",
    "expected": ["baking recipe"],
    "not_labels_explained": {
      "cooking chemistry": "No explanation of the science behind creaming or emulsion.",
      "food blog": "The imperative procedural tone is pure recipe, not a blog's conversational style.",
      "kitchen equipment review": "Equipment (oven, bowl) is mentioned instrumentally, not reviewed.",
      "nutrition information": "No calorie or nutrient data is provided."
    }
  },
  {
    "text": "The proposed amendment would require all political advertisements to disclose their funding sources within the ad itself, not just in FEC filings.",
    "labels": ["campaign finance reform", "transparency legislation", "media regulation", "first amendment issue", "election interference"],
    "type": "multi-label",
    "expected": ["campaign finance reform", "transparency legislation"],
    "not_labels_explained": {
      "media regulation": "The regulation targets political ads specifically, not media broadly.",
      "first amendment issue": "Disclosure requirements are described neutrally, not framed as a speech issue.",
      "election interference": "This is a proposed domestic regulation, not interference."
    }
  },
  {
    "text": "I scored a 720 on the GMAT. Is that good enough for Harvard Business School, or should I retake it?",
    "labels": ["admissions question", "test score evaluation", "academic anxiety", "humble brag", "career planning"],
    "type": "multi-label",
    "expected": ["admissions question", "test score evaluation"],
    "not_labels_explained": {
      "academic anxiety": "The question is practical, not anxious in tone.",
      "humble brag": "A 720 is strong but the question appears genuinely uncertain, not showing off.",
      "career planning": "The question is about admissions, not career trajectory."
    }
  },
  {
    "text": "The building's facade combines Art Deco geometric patterns with Brutalist concrete massing, creating a jarring but oddly compelling visual tension.",
    "labels": ["architecture criticism", "building description", "design theory", "urban planning", "real estate listing"],
    "type": "multi-label",
    "expected": ["architecture criticism", "building description"],
    "not_labels_explained": {
      "design theory": "The observation is specific to one building, not a theoretical framework.",
      "urban planning": "The building's relationship to its surroundings is not discussed.",
      "real estate listing": "Aesthetic evaluation, not a sales description."
    }
  },
  {
    "text": "The autonomous vehicle navigated the construction zone without incident, correctly identifying lane closures, workers, and temporary signage in real time.",
    "labels": ["autonomous driving test", "traffic report", "safety certification", "technology demonstration", "accident report"],
    "type": "multi-label",
    "expected": ["autonomous driving test", "technology demonstration"],
    "not_labels_explained": {
      "traffic report": "This reports on the vehicle's capabilities, not traffic conditions.",
      "safety certification": "A successful test is not the same as formal certification.",
      "accident report": "No accident occurred; the vehicle navigated successfully."
    }
  },
  {
    "text": "Dear Senator, as a veteran who served two tours in Afghanistan, I urge you to vote against the proposed cuts to VA healthcare. Our servicemembers deserve better.",
    "labels": ["constituent letter", "veteran advocacy", "political lobbying", "military memoir", "healthcare policy"],
    "type": "multi-label",
    "expected": ["constituent letter", "veteran advocacy"],
    "not_labels_explained": {
      "political lobbying": "This is a personal constituent letter, not organized professional lobbying.",
      "military memoir": "Military service is mentioned as credentials, not narrated as memoir.",
      "healthcare policy": "The letter opposes cuts but doesn't analyze or propose policy."
    }
  },
  {
    "text": "The meeting is at 3pm but honestly I'd rather stab myself with a fork than sit through another two-hour update on the Q3 roadmap.",
    "labels": ["workplace frustration", "meeting complaint", "self-harm", "humor", "schedule notification"],
    "type": "multi-label",
    "expected": ["workplace frustration", "meeting complaint", "humor"],
    "not_labels_explained": {
      "self-harm": "The 'stab myself with a fork' is hyperbolic frustration, not genuine self-harm ideation.",
      "schedule notification": "The 3pm detail is incidental context, not the purpose of the message."
    }
  },
  {
    "text": "We are pleased to announce that Dr. Sarah Chen has been appointed as the new Chief Medical Officer, effective immediately. Dr. Chen brings 20 years of experience in infectious disease research.",
    "labels": ["executive appointment", "press release", "medical news", "career achievement", "organizational change"],
    "type": "multi-label",
    "expected": ["executive appointment", "press release", "organizational change"],
    "not_labels_explained": {
      "medical news": "This is a personnel announcement, not news about medicine or health.",
      "career achievement": "The focus is on the organization's decision, not Dr. Chen's personal career narrative."
    }
  },
  {
    "text": "I love how my neighbor plays drums at midnight. Really adds to the ambiance of trying to sleep before a 6am shift.",
    "labels": ["noise complaint", "sarcasm", "neighbor appreciation", "sleep deprivation", "music review"],
    "type": "multi-label",
    "expected": ["noise complaint", "sarcasm", "sleep deprivation"],
    "not_labels_explained": {
      "neighbor appreciation": "'I love how' is clearly sarcastic given the context of midnight drums and a 6am shift.",
      "music review": "No evaluation of musical quality is intended."
    }
  },
  {
    "text": "The model achieves 94.2% accuracy on the test set, but further investigation reveals it memorized the training data — performance drops to 61% on out-of-distribution samples.",
    "labels": ["machine learning evaluation", "overfitting diagnosis", "research breakthrough", "benchmark result", "technical criticism"],
    "type": "multi-label",
    "expected": ["machine learning evaluation", "overfitting diagnosis", "technical criticism"],
    "not_labels_explained": {
      "research breakthrough": "The finding reveals a failure (overfitting), not a breakthrough.",
      "benchmark result": "The 94.2% is explicitly shown to be misleading due to overfitting."
    }
  }
]