dataset.json

[
    {
        "instruction": "What is Tone can be?",
        "input": "",
        "output": "Tone can be: critical, appreciative, neutral, sarcastic, analytical",
        "text": "### Instruction:\nWhat is Tone can be?\n\n### Response:\nTone can be: critical, appreciative, neutral, sarcastic, analytical"
    },
    {
        "instruction": "What is Random Forest?",
        "input": "",
        "output": "Random Forest: ensemble of decision trees, reduces overfitting",
        "text": "### Instruction:\nWhat is Random Forest?\n\n### Response:\nRandom Forest: ensemble of decision trees, reduces overfitting"
    },
    {
        "instruction": "What is Complementary pairs?",
        "input": "",
        "output": "Complementary pairs: >, < and >=, <=",
        "text": "### Instruction:\nWhat is Complementary pairs?\n\n### Response:\nComplementary pairs: >, < and >=, <="
    },
    {
        "instruction": "What is Logistic Regression?",
        "input": "",
        "output": "Logistic Regression: binary classification, outputs probability",
        "text": "### Instruction:\nWhat is Logistic Regression?\n\n### Response:\nLogistic Regression: binary classification, outputs probability"
    },
    {
        "instruction": "What is the formula for --- Time, Speed and Distance ---?",
        "input": "",
        "output": "Distance = Speed x Time",
        "text": "### Instruction:\nWhat is the formula for --- Time, Speed and Distance ---?\n\n### Response:\nDistance = Speed x Time"
    },
    {
        "instruction": "What is the formula for Venn diagrams help visualize syllogism problems?",
        "input": "",
        "output": "Father's father = Paternal grandfather",
        "text": "### Instruction:\nWhat is the formula for Venn diagrams help visualize syllogism problems?\n\n### Response:\nFather's father = Paternal grandfather"
    },
    {
        "instruction": "What is Divisibility by 11?",
        "input": "",
        "output": "Divisibility by 11: difference of alternate digit sums divisible by 11",
        "text": "### Instruction:\nWhat is Divisibility by 11?\n\n### Response:\nDivisibility by 11: difference of alternate digit sums divisible by 11"
    },
    {
        "instruction": "What is RBI established?",
        "input": "",
        "output": "RBI established: 1 April 1935",
        "text": "### Instruction:\nWhat is RBI established?\n\n### Response:\nRBI established: 1 April 1935"
    },
    {
        "instruction": "What is Private sector?",
        "input": "",
        "output": "Private sector: majority private owned (HDFC, ICICI, Axis)",
        "text": "### Instruction:\nWhat is Private sector?\n\n### Response:\nPrivate sector: majority private owned (HDFC, ICICI, Axis)"
    },
    {
        "instruction": "What is the formula for --- Indian Economy ---?",
        "input": "",
        "output": "NNP: Net National Product = GNP - Depreciation",
        "text": "### Instruction:\nWhat is the formula for --- Indian Economy ---?\n\n### Response:\nNNP: Net National Product = GNP - Depreciation"
    },
    {
        "instruction": "What is IMF?",
        "input": "",
        "output": "IMF: International Monetary Fund - maintains global financial stability",
        "text": "### Instruction:\nWhat is IMF?\n\n### Response:\nIMF: International Monetary Fund - maintains global financial stability"
    },
    {
        "instruction": "What is Text Classification?",
        "input": "",
        "output": "Text Classification: assign categories to text",
        "text": "### Instruction:\nWhat is Text Classification?\n\n### Response:\nText Classification: assign categories to text"
    },
    {
        "instruction": "What is the formula for --- Time and Work ---?",
        "input": "",
        "output": "If A and B together complete work in n days, combined 1 day work = 1/n",
        "text": "### Instruction:\nWhat is the formula for --- Time and Work ---?\n\n### Response:\nIf A and B together complete work in n days, combined 1 day work = 1/n"
    },
    {
        "instruction": "What is GRU (Gated Recurrent Unit)?",
        "input": "",
        "output": "GRU (Gated Recurrent Unit): simpler than LSTM, similar performance",
        "text": "### Instruction:\nWhat is GRU (Gated Recurrent Unit)?\n\n### Response:\nGRU (Gated Recurrent Unit): simpler than LSTM, similar performance"
    },
    {
        "instruction": "What is Bias?",
        "input": "",
        "output": "Bias: error from wrong assumptions in learning algorithm",
        "text": "### Instruction:\nWhat is Bias?\n\n### Response:\nBias: error from wrong assumptions in learning algorithm"
    },
    {
        "instruction": "What is the formula for Relative speed (opposite direction) = S1 + S2?",
        "input": "",
        "output": "If A completes work in n days, A's 1 day work = 1/n",
        "text": "### Instruction:\nWhat is the formula for Relative speed (opposite direction) = S1 + S2?\n\n### Response:\nIf A completes work in n days, A's 1 day work = 1/n"
    },
    {
        "instruction": "What is Confusion Matrix?",
        "input": "",
        "output": "Confusion Matrix: TP, TN, FP, FN",
        "text": "### Instruction:\nWhat is Confusion Matrix?\n\n### Response:\nConfusion Matrix: TP, TN, FP, FN"
    },
    {
        "instruction": "What is Divisibility by 5?",
        "input": "",
        "output": "Divisibility by 5: last digit is 0 or 5",
        "text": "### Instruction:\nWhat is Divisibility by 5?\n\n### Response:\nDivisibility by 5: last digit is 0 or 5"
    },
    {
        "instruction": "What is Square series?",
        "input": "",
        "output": "Square series: 1,4,9,16,25...",
        "text": "### Instruction:\nWhat is Square series?\n\n### Response:\nSquare series: 1,4,9,16,25..."
    },
    {
        "instruction": "What are the key concepts in Coding-Decoding?",
        "input": "",
        "output": "Letter coding: A=1, B=2, ..., Z=26 (or reverse Z=1)\nShift coding: each letter shifted by fixed number\nSymbol coding: letters replaced by symbols\nNumber coding: words replaced by numbers\nMirror image: A corresponds to Z, B to Y, C to X, etc.",
        "text": "### Instruction:\nWhat are the key concepts in Coding-Decoding?\n\n### Response:\nLetter coding: A=1, B=2, ..., Z=26 (or reverse Z=1)\nShift coding: each letter shifted by fixed number\nSymbol coding: letters replaced by symbols\nNumber coding: words replaced by numbers\nMirror image: A corresponds to Z, B to Y, C to X, etc."
    },
    {
        "instruction": "What is Shortest distance?",
        "input": "",
        "output": "Shortest distance: Pythagoras theorem for perpendicular paths",
        "text": "### Instruction:\nWhat is Shortest distance?\n\n### Response:\nShortest distance: Pythagoras theorem for perpendicular paths"
    },
    {
        "instruction": "What is Pradhan Mantri Suraksha Bima Yojana (PMSBY)?",
        "input": "",
        "output": "Pradhan Mantri Suraksha Bima Yojana (PMSBY): accident insurance",
        "text": "### Instruction:\nWhat is Pradhan Mantri Suraksha Bima Yojana (PMSBY)?\n\n### Response:\nPradhan Mantri Suraksha Bima Yojana (PMSBY): accident insurance"
    },
    {
        "instruction": "What is Backpropagation?",
        "input": "",
        "output": "Backpropagation: computes gradients by chain rule, flowing backward through network",
        "text": "### Instruction:\nWhat is Backpropagation?\n\n### Response:\nBackpropagation: computes gradients by chain rule, flowing backward through network"
    },
    {
        "instruction": "What are the key concepts in Algorithms?",
        "input": "",
        "output": "Linear Regression: predict continuous values, fit line to data\n  - Loss: Mean Squared Error (MSE)\n  - y = mx + c\n\nLogistic Regression: binary classification, outputs probability\n  - Uses sigmoid function: 1/(1+e^-x)\n  - Loss: Binary Cross-Entropy\n\nDecision Tree: tree structure, splits data on features\n  - Gini Impurity: measures split quality\n  - Information Gain: entropy-based measure\n\nRandom Forest: ensemble of decision trees, reduces overfitting\n  - Bagging: train each tree on random subset of data\n  - Feature importance: how much each feature contributes\n\nSVM: Support Vector Machine, finds",
        "text": "### Instruction:\nWhat are the key concepts in Algorithms?\n\n### Response:\nLinear Regression: predict continuous values, fit line to data\n  - Loss: Mean Squared Error (MSE)\n  - y = mx + c\n\nLogistic Regression: binary classification, outputs probability\n  - Uses sigmoid function: 1/(1+e^-x)\n  - Loss: Binary Cross-Entropy\n\nDecision Tree: tree structure, splits data on features\n  - Gini Impurity: measures split quality\n  - Information Gain: entropy-based measure\n\nRandom Forest: ensemble of decision trees, reduces overfitting\n  - Bagging: train each tree on random subset of data\n  - Feature importance: how much each feature contributes\n\nSVM: Support Vector Machine, finds"
    },
    {
        "instruction": "What are the key concepts in Ratio and Proportion?",
        "input": "",
        "output": "Ratio a:b means a/b\nIf a:b = c:d then ad = bc (cross multiplication)\nCompound ratio of a:b and c:d is ac:bd\nDuplicate ratio of a:b is a^2:b^2\nSub-duplicate ratio of a:b is sqrt(a):sqrt(b)\nIf A:B = 2:3 and B:C = 4:5, then A:B:C = 8:12:15\nProportion: a:b = c:d means a, b, c, d are in proportion\nMean proportion of a and b = sqrt(ab)",
        "text": "### Instruction:\nWhat are the key concepts in Ratio and Proportion?\n\n### Response:\nRatio a:b means a/b\nIf a:b = c:d then ad = bc (cross multiplication)\nCompound ratio of a:b and c:d is ac:bd\nDuplicate ratio of a:b is a^2:b^2\nSub-duplicate ratio of a:b is sqrt(a):sqrt(b)\nIf A:B = 2:3 and B:C = 4:5, then A:B:C = 8:12:15\nProportion: a:b = c:d means a, b, c, d are in proportion\nMean proportion of a and b = sqrt(ab)"
    },
    {
        "instruction": "What is the formula for If two items sold at same price, one at x% profit and one at x% loss, overall loss% = x^2/100?",
        "input": "",
        "output": "Discount% = (Discount / Marked Price) x 100",
        "text": "### Instruction:\nWhat is the formula for If two items sold at same price, one at x% profit and one at x% loss, overall loss% = x^2/100?\n\n### Response:\nDiscount% = (Discount / Marked Price) x 100"
    },
    {
        "instruction": "What is Cylinder?",
        "input": "",
        "output": "Cylinder: Volume = pi x r^2 x h, Curved Surface Area = 2 x pi x r x h",
        "text": "### Instruction:\nWhat is Cylinder?\n\n### Response:\nCylinder: Volume = pi x r^2 x h, Curved Surface Area = 2 x pi x r x h"
    },
    {
        "instruction": "What are the key concepts in Important Financial Institutions?",
        "input": "",
        "output": "SEBI: Securities and Exchange Board of India - regulates stock markets\nIRDA/IRDAI: Insurance Regulatory and Development Authority of India\nPFRDA: Pension Fund Regulatory and Development Authority\nNABARD: National Bank for Agriculture and Rural Development\nNHB: National Housing Bank - regulates housing finance companies\nSIDBI: Small Industries Development Bank of India\nEXIM Bank: Export-Import Bank of India\nWorld Bank: provides loans for development projects\nIMF: International Monetary Fund - maintains global financial stability\nADB: Asian Development Bank - promotes development in Asia\nAIIB: A",
        "text": "### Instruction:\nWhat are the key concepts in Important Financial Institutions?\n\n### Response:\nSEBI: Securities and Exchange Board of India - regulates stock markets\nIRDA/IRDAI: Insurance Regulatory and Development Authority of India\nPFRDA: Pension Fund Regulatory and Development Authority\nNABARD: National Bank for Agriculture and Rural Development\nNHB: National Housing Bank - regulates housing finance companies\nSIDBI: Small Industries Development Bank of India\nEXIM Bank: Export-Import Bank of India\nWorld Bank: provides loans for development projects\nIMF: International Monetary Fund - maintains global financial stability\nADB: Asian Development Bank - promotes development in Asia\nAIIB: A"
    },
    {
        "instruction": "What is Moving opposite direction?",
        "input": "",
        "output": "Moving opposite direction: Time = (L1+L2) / (S1+S2)",
        "text": "### Instruction:\nWhat is Moving opposite direction?\n\n### Response:\nMoving opposite direction: Time = (L1+L2) / (S1+S2)"
    },
    {
        "instruction": "What are the key concepts in Emotion Detector?",
        "input": "",
        "output": "What: Real-time emotion detection with recommendations\nTech: TensorFlow, OpenCV, PyQt5\nKey Features:\n  - CNN for facial emotion recognition (7 emotions)\n  - PyQt5 desktop GUI\n  - Emotion-based content recommender (100+ suggestions)\n  - Sub-second inference",
        "text": "### Instruction:\nWhat are the key concepts in Emotion Detector?\n\n### Response:\nWhat: Real-time emotion detection with recommendations\nTech: TensorFlow, OpenCV, PyQt5\nKey Features:\n  - CNN for facial emotion recognition (7 emotions)\n  - PyQt5 desktop GUI\n  - Emotion-based content recommender (100+ suggestions)\n  - Sub-second inference"
    },
    {
        "instruction": "What is Matplotlib/Seaborn?",
        "input": "",
        "output": "Matplotlib/Seaborn: data visualization",
        "text": "### Instruction:\nWhat is Matplotlib/Seaborn?\n\n### Response:\nMatplotlib/Seaborn: data visualization"
    },
    {
        "instruction": "What is Feature extraction?",
        "input": "",
        "output": "Feature extraction: freeze pretrained layers, train only final layers",
        "text": "### Instruction:\nWhat is Feature extraction?\n\n### Response:\nFeature extraction: freeze pretrained layers, train only final layers"
    },
    {
        "instruction": "What is Valid padding?",
        "input": "",
        "output": "Valid padding: no padding, output smaller",
        "text": "### Instruction:\nWhat is Valid padding?\n\n### Response:\nValid padding: no padding, output smaller"
    },
    {
        "instruction": "What is the formula for Father's father = Paternal grandfather?",
        "input": "",
        "output": "Husband's mother = Mother-in-law",
        "text": "### Instruction:\nWhat is the formula for Father's father = Paternal grandfather?\n\n### Response:\nHusband's mother = Mother-in-law"
    },
    {
        "instruction": "What is TensorFlow/Keras?",
        "input": "",
        "output": "TensorFlow/Keras: deep learning framework",
        "text": "### Instruction:\nWhat is TensorFlow/Keras?\n\n### Response:\nTensorFlow/Keras: deep learning framework"
    },
    {
        "instruction": "What is SECTION 5?",
        "input": "",
        "output": "SECTION 5: GENERAL AWARENESS / CURRENT AFFAIRS (STATIC GK)",
        "text": "### Instruction:\nWhat is SECTION 5?\n\n### Response:\nSECTION 5: GENERAL AWARENESS / CURRENT AFFAIRS (STATIC GK)"
    },
    {
        "instruction": "What are the key concepts in RNN and Transformers?",
        "input": "",
        "output": "RNN (Recurrent Neural Network): processes sequences, has memory\n  - Problem: vanishing gradient for long sequences\nLSTM (Long Short-Term Memory): solves vanishing gradient\n  - Cell state, forget gate, input gate, output gate\nGRU (Gated Recurrent Unit): simpler than LSTM, similar performance\n\nTransformer: attention-based architecture, no recurrence\n  - Self-attention: each token attends to all other tokens\n  - Attention formula: softmax(QK^T / sqrt(d_k)) x V\n  - Q: Query, K: Key, V: Value\n  - Multi-head attention: run attention multiple times in parallel\n  - Positional encoding: adds position i",
        "text": "### Instruction:\nWhat are the key concepts in RNN and Transformers?\n\n### Response:\nRNN (Recurrent Neural Network): processes sequences, has memory\n  - Problem: vanishing gradient for long sequences\nLSTM (Long Short-Term Memory): solves vanishing gradient\n  - Cell state, forget gate, input gate, output gate\nGRU (Gated Recurrent Unit): simpler than LSTM, similar performance\n\nTransformer: attention-based architecture, no recurrence\n  - Self-attention: each token attends to all other tokens\n  - Attention formula: softmax(QK^T / sqrt(d_k)) x V\n  - Q: Query, K: Key, V: Value\n  - Multi-head attention: run attention multiple times in parallel\n  - Positional encoding: adds position i"
    },
    {
        "instruction": "A pipe fills a tank in 6 hours. Another pipe empties it in 10 hours. If both are open, in how many hours will the tank fill?",
        "output": "Fill rate = 1/6, Empty rate = 1/10. Net rate = 1/6 - 1/10 = 5/30 - 3/30 = 2/30 = 1/15. Tank fills in 15 hours.",
        "text": "### Instruction:\nA pipe fills a tank in 6 hours. Another pipe empties it in 10 hours. If both are open, in how many hours will the tank fill?\n\n### Response:\nFill rate = 1/6, Empty rate = 1/10. Net rate = 1/6 - 1/10 = 5/30 - 3/30 = 2/30 = 1/15. Tank fills in 15 hours."
    },
    {
        "instruction": "What number comes next in the series: 2, 6, 12, 20, 30, ?",
        "output": "Pattern: 1x2, 2x3, 3x4, 4x5, 5x6, next = 6x7 = 42",
        "text": "### Instruction:\nWhat number comes next in the series: 2, 6, 12, 20, 30, ?\n\n### Response:\nPattern: 1x2, 2x3, 3x4, 4x5, 5x6, next = 6x7 = 42"
    },
    {
        "instruction": "What are the key concepts in Important Days?",
        "input": "",
        "output": "26 January: Republic Day\n15 August: Independence Day\n2 October: Gandhi Jayanti\n14 November: Children's Day (Nehru's birthday)\n14 April: Ambedkar Jayanti\n1 May: Labour Day / Maharashtra Day / Gujarat Day\n5 September: Teachers Day (Radhakrishnan's birthday)\n1 April: RBI Foundation Day\n18 April: World Heritage Day\n29 July: National Tiger Day\n\n================================================================================\nSECTION 6: COMPUTER AWARENESS (FOR BANKING EXAMS)\n================================================================================",
        "text": "### Instruction:\nWhat are the key concepts in Important Days?\n\n### Response:\n26 January: Republic Day\n15 August: Independence Day\n2 October: Gandhi Jayanti\n14 November: Children's Day (Nehru's birthday)\n14 April: Ambedkar Jayanti\n1 May: Labour Day / Maharashtra Day / Gujarat Day\n5 September: Teachers Day (Radhakrishnan's birthday)\n1 April: RBI Foundation Day\n18 April: World Heritage Day\n29 July: National Tiger Day\n\n================================================================================\nSECTION 6: COMPUTER AWARENESS (FOR BANKING EXAMS)\n================================================================================"
    },
    {
        "instruction": "What is the significance/value of Article 14?",
        "input": "",
        "output": "Article 14: Equality before law",
        "text": "### Instruction:\nWhat is the significance/value of Article 14?\n\n### Response:\nArticle 14: Equality before law"
    },
    {
        "instruction": "What is LSTM (Long Short-Term Memory)?",
        "input": "",
        "output": "LSTM (Long Short-Term Memory): solves vanishing gradient",
        "text": "### Instruction:\nWhat is LSTM (Long Short-Term Memory)?\n\n### Response:\nLSTM (Long Short-Term Memory): solves vanishing gradient"
    },
    {
        "instruction": "What is No A is B?",
        "input": "",
        "output": "No A is B: A and B have no overlap",
        "text": "### Instruction:\nWhat is No A is B?\n\n### Response:\nNo A is B: A and B have no overlap"
    },
    {
        "instruction": "What is the formula for Cost Price = Selling Price / (1 + Profit%/100)?",
        "input": "",
        "output": "Discount = Marked Price - Selling Price",
        "text": "### Instruction:\nWhat is the formula for Cost Price = Selling Price / (1 + Profit%/100)?\n\n### Response:\nDiscount = Marked Price - Selling Price"
    },
    {
        "instruction": "What are the key concepts in Python for ML?",
        "input": "",
        "output": "NumPy: numerical computing, arrays, matrix operations\n  - np.array(), np.reshape(), np.dot(), np.mean(), np.std()\n  - Broadcasting: operations on different shaped arrays\nPandas: data manipulation and analysis\n  - pd.read_csv(), df.head(), df.describe(), df.dropna(), df.fillna()\n  - df.groupby(), df.merge(), df.pivot_table()\nMatplotlib/Seaborn: data visualization\nScikit-learn: ML algorithms, preprocessing, evaluation\n  - train_test_split, StandardScaler, LabelEncoder\n  - KMeans, RandomForestClassifier, LinearRegression\nTensorFlow/Keras: deep learning framework\n  - model.fit(), model.predict(),",
        "text": "### Instruction:\nWhat are the key concepts in Python for ML?\n\n### Response:\nNumPy: numerical computing, arrays, matrix operations\n  - np.array(), np.reshape(), np.dot(), np.mean(), np.std()\n  - Broadcasting: operations on different shaped arrays\nPandas: data manipulation and analysis\n  - pd.read_csv(), df.head(), df.describe(), df.dropna(), df.fillna()\n  - df.groupby(), df.merge(), df.pivot_table()\nMatplotlib/Seaborn: data visualization\nScikit-learn: ML algorithms, preprocessing, evaluation\n  - train_test_split, StandardScaler, LabelEncoder\n  - KMeans, RandomForestClassifier, LinearRegression\nTensorFlow/Keras: deep learning framework\n  - model.fit(), model.predict(),"
    },
    {
        "instruction": "What is the formula for If price increases by r%, new price = original x (1 + r/100)?",
        "input": "",
        "output": "Percentage decrease = (decrease / original) x 100",
        "text": "### Instruction:\nWhat is the formula for If price increases by r%, new price = original x (1 + r/100)?\n\n### Response:\nPercentage decrease = (decrease / original) x 100"
    },
    {
        "instruction": "What is When compounded half-yearly?",
        "input": "",
        "output": "When compounded half-yearly: Rate = R/2, Time = 2T",
        "text": "### Instruction:\nWhat is When compounded half-yearly?\n\n### Response:\nWhen compounded half-yearly: Rate = R/2, Time = 2T"
    },
    {
        "instruction": "What is the formula for --- Averages ---?",
        "input": "",
        "output": "Sum = Average x Number",
        "text": "### Instruction:\nWhat is the formula for --- Averages ---?\n\n### Response:\nSum = Average x Number"
    },
    {
        "instruction": "What is Arithmetic series?",
        "input": "",
        "output": "Arithmetic series: difference is constant (2,4,6,8...)",
        "text": "### Instruction:\nWhat is Arithmetic series?\n\n### Response:\nArithmetic series: difference is constant (2,4,6,8...)"
    },
    {
        "instruction": "What is the significance/value of Article 1?",
        "input": "",
        "output": "Article 1: India shall be a Union of States",
        "text": "### Instruction:\nWhat is the significance/value of Article 1?\n\n### Response:\nArticle 1: India shall be a Union of States"
    },
    {
        "instruction": "What is RBI Governor?",
        "input": "",
        "output": "RBI Governor: appointed by Central Government",
        "text": "### Instruction:\nWhat is RBI Governor?\n\n### Response:\nRBI Governor: appointed by Central Government"
    },
    {
        "instruction": "What are the key concepts in Computer Vision?",
        "input": "",
        "output": "Image preprocessing: resize, normalize, augment\nData augmentation: flip, rotate, zoom, shift, brightness change\n  - Used in Drive-Awake training\nTransfer Learning: use pre-trained model weights on new task\n  - Feature extraction: freeze pretrained layers, train only final layers\n  - Fine-tuning: unfreeze and train all layers with small learning rate\nObject Detection: locate and classify objects in image\n  - YOLO: You Only Look Once, real-time detection\n  - R-CNN family: region-based detection\n\nOpenCV:\n  - cv2.imread(): read image\n  - cv2.cvtColor(): convert color space (BGR to RGB, BGR to Gray",
        "text": "### Instruction:\nWhat are the key concepts in Computer Vision?\n\n### Response:\nImage preprocessing: resize, normalize, augment\nData augmentation: flip, rotate, zoom, shift, brightness change\n  - Used in Drive-Awake training\nTransfer Learning: use pre-trained model weights on new task\n  - Feature extraction: freeze pretrained layers, train only final layers\n  - Fine-tuning: unfreeze and train all layers with small learning rate\nObject Detection: locate and classify objects in image\n  - YOLO: You Only Look Once, real-time detection\n  - R-CNN family: region-based detection\n\nOpenCV:\n  - cv2.imread(): read image\n  - cv2.cvtColor(): convert color space (BGR to RGB, BGR to Gray"
    },
    {
        "instruction": "What is the formula for If A > B > C, then A > C is definitely true?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for If A > B > C, then A > C is definitely true?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is Article 15?",
        "input": "",
        "output": "Article 15: Prohibition of discrimination",
        "text": "### Instruction:\nWhat is Article 15?\n\n### Response:\nArticle 15: Prohibition of discrimination"
    },
    {
        "instruction": "What is the significance/value of Article 44?",
        "input": "",
        "output": "Article 44: Uniform Civil Code (Directive Principle)",
        "text": "### Instruction:\nWhat is the significance/value of Article 44?\n\n### Response:\nArticle 44: Uniform Civil Code (Directive Principle)"
    },
    {
        "instruction": "What is NITI Aayog?",
        "input": "",
        "output": "NITI Aayog: replaced Planning Commission in 2015",
        "text": "### Instruction:\nWhat is NITI Aayog?\n\n### Response:\nNITI Aayog: replaced Planning Commission in 2015"
    },
    {
        "instruction": "What is Reverse Repo Rate?",
        "input": "",
        "output": "Reverse Repo Rate: rate at which RBI borrows from banks",
        "text": "### Instruction:\nWhat is Reverse Repo Rate?\n\n### Response:\nReverse Repo Rate: rate at which RBI borrows from banks"
    },
    {
        "instruction": "What is MSE (Mean Squared Error)?",
        "input": "",
        "output": "MSE (Mean Squared Error): average squared difference",
        "text": "### Instruction:\nWhat is MSE (Mean Squared Error)?\n\n### Response:\nMSE (Mean Squared Error): average squared difference"
    },
    {
        "instruction": "What is cv2.cvtColor()?",
        "input": "",
        "output": "cv2.cvtColor(): convert color space (BGR to RGB, BGR to Gray)",
        "text": "### Instruction:\nWhat is cv2.cvtColor()?\n\n### Response:\ncv2.cvtColor(): convert color space (BGR to RGB, BGR to Gray)"
    },
    {
        "instruction": "What is Integers?",
        "input": "",
        "output": "Integers: ...-3, -2, -1, 0, 1, 2, 3...",
        "text": "### Instruction:\nWhat is Integers?\n\n### Response:\nIntegers: ...-3, -2, -1, 0, 1, 2, 3..."
    },
    {
        "instruction": "What is Common confusions?",
        "input": "",
        "output": "Common confusions: affect/effect, than/then, its/it's, there/their/they're",
        "text": "### Instruction:\nWhat is Common confusions?\n\n### Response:\nCommon confusions: affect/effect, than/then, its/it's, there/their/they're"
    },
    {
        "instruction": "What is Embedding?",
        "input": "",
        "output": "Embedding: mapping tokens to dense vectors (Word2Vec, GloVe, FastText)",
        "text": "### Instruction:\nWhat is Embedding?\n\n### Response:\nEmbedding: mapping tokens to dense vectors (Word2Vec, GloVe, FastText)"
    },
    {
        "instruction": "What is Transformer?",
        "input": "",
        "output": "Transformer: attention-based architecture, no recurrence",
        "text": "### Instruction:\nWhat is Transformer?\n\n### Response:\nTransformer: attention-based architecture, no recurrence"
    },
    {
        "instruction": "What is the formula for END OF NOTES?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for END OF NOTES?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What are the key concepts in Percentage?",
        "input": "",
        "output": "Percentage means per hundred. x% = x/100\nIf price increases by r%, new price = original x (1 + r/100)\nIf price decreases by r%, new price = original x (1 - r/100)\nPercentage increase = (increase / original) x 100\nPercentage decrease = (decrease / original) x 100\nIf A is x% more than B, B is less than A by: (x / 100+x) x 100 percent\nIf A is x% less than B, B is more than A by: (x / 100-x) x 100 percent",
        "text": "### Instruction:\nWhat are the key concepts in Percentage?\n\n### Response:\nPercentage means per hundred. x% = x/100\nIf price increases by r%, new price = original x (1 + r/100)\nIf price decreases by r%, new price = original x (1 - r/100)\nPercentage increase = (increase / original) x 100\nPercentage decrease = (decrease / original) x 100\nIf A is x% more than B, B is less than A by: (x / 100+x) x 100 percent\nIf A is x% less than B, B is more than A by: (x / 100-x) x 100 percent"
    },
    {
        "instruction": "What is the significance/value of Article 12?",
        "input": "",
        "output": "Article 12: Definition of State",
        "text": "### Instruction:\nWhat is the significance/value of Article 12?\n\n### Response:\nArticle 12: Definition of State"
    },
    {
        "instruction": "What is Public sector?",
        "input": "",
        "output": "Public sector: majority government owned (SBI, BOB, PNB)",
        "text": "### Instruction:\nWhat is Public sector?\n\n### Response:\nPublic sector: majority government owned (SBI, BOB, PNB)"
    },
    {
        "instruction": "What is Small Finance Banks?",
        "input": "",
        "output": "Small Finance Banks: serve small borrowers and unserved sections",
        "text": "### Instruction:\nWhat is Small Finance Banks?\n\n### Response:\nSmall Finance Banks: serve small borrowers and unserved sections"
    },
    {
        "instruction": "What are the key concepts in Drive-Awake (Drowsiness Detection)?",
        "input": "",
        "output": "What: Real-time driver drowsiness detection system\nTech: Python, TensorFlow, FastAPI, OpenCV, pygame, Docker\nKey Features:\n  - CNN model for eye state classification (open/closed)\n  - Haar Cascade for face detection\n  - Eye Aspect Ratio (EAR) for blink pattern analysis\n  - Audio alarm when drowsiness detected\n  - 92%+ accuracy, real-time at 30+ FPS\n  - FastAPI for REST API deployment\n  - Docker containerization\nImpact: Prevents road accidents from driver fatigue\nBugs to fix: duplicate requirements, Dockerfile OS mismatch, float() on array",
        "text": "### Instruction:\nWhat are the key concepts in Drive-Awake (Drowsiness Detection)?\n\n### Response:\nWhat: Real-time driver drowsiness detection system\nTech: Python, TensorFlow, FastAPI, OpenCV, pygame, Docker\nKey Features:\n  - CNN model for eye state classification (open/closed)\n  - Haar Cascade for face detection\n  - Eye Aspect Ratio (EAR) for blink pattern analysis\n  - Audio alarm when drowsiness detected\n  - 92%+ accuracy, real-time at 30+ FPS\n  - FastAPI for REST API deployment\n  - Docker containerization\nImpact: Prevents road accidents from driver fatigue\nBugs to fix: duplicate requirements, Dockerfile OS mismatch, float() on array"
    },
    {
        "instruction": "What are the key concepts in Fill in the Blanks?",
        "input": "",
        "output": "Check grammar: noun/verb/adjective/adverb required\nCheck meaning: word must fit context\nCommon confusions: affect/effect, than/then, its/it's, there/their/they're\n\n================================================================================\nSECTION 4: BANKING AND FINANCIAL AWARENESS\n================================================================================",
        "text": "### Instruction:\nWhat are the key concepts in Fill in the Blanks?\n\n### Response:\nCheck grammar: noun/verb/adjective/adverb required\nCheck meaning: word must fit context\nCommon confusions: affect/effect, than/then, its/it's, there/their/they're\n\n================================================================================\nSECTION 4: BANKING AND FINANCIAL AWARENESS\n================================================================================"
    },
    {
        "instruction": "What is Foreign Key?",
        "input": "",
        "output": "Foreign Key: links two tables",
        "text": "### Instruction:\nWhat is Foreign Key?\n\n### Response:\nForeign Key: links two tables"
    },
    {
        "instruction": "What is the formula for --- Monetary Policy Tools ---?",
        "input": "",
        "output": "- Increase in repo rate = credit becomes expensive = inflation control",
        "text": "### Instruction:\nWhat is the formula for --- Monetary Policy Tools ---?\n\n### Response:\n- Increase in repo rate = credit becomes expensive = inflation control"
    },
    {
        "instruction": "What is ROC-AUC?",
        "input": "",
        "output": "ROC-AUC: area under ROC curve, higher is better",
        "text": "### Instruction:\nWhat is ROC-AUC?\n\n### Response:\nROC-AUC: area under ROC curve, higher is better"
    },
    {
        "instruction": "What is Pipe and Cistern?",
        "input": "",
        "output": "Pipe and Cistern: inlet pipes fill, outlet pipes empty",
        "text": "### Instruction:\nWhat is Pipe and Cistern?\n\n### Response:\nPipe and Cistern: inlet pipes fill, outlet pipes empty"
    },
    {
        "instruction": "What is R-squared?",
        "input": "",
        "output": "R-squared: proportion of variance explained by model (0 to 1)",
        "text": "### Instruction:\nWhat is R-squared?\n\n### Response:\nR-squared: proportion of variance explained by model (0 to 1)"
    },
    {
        "instruction": "What is Find closing sentence?",
        "input": "",
        "output": "Find closing sentence: concludes or summarizes",
        "text": "### Instruction:\nWhat is Find closing sentence?\n\n### Response:\nFind closing sentence: concludes or summarizes"
    },
    {
        "instruction": "What is Kishore?",
        "input": "",
        "output": "Kishore: 50,000 to 5 lakh",
        "text": "### Instruction:\nWhat is Kishore?\n\n### Response:\nKishore: 50,000 to 5 lakh"
    },
    {
        "instruction": "What is Positional encoding?",
        "input": "",
        "output": "Positional encoding: adds position information",
        "text": "### Instruction:\nWhat is Positional encoding?\n\n### Response:\nPositional encoding: adds position information"
    },
    {
        "instruction": "What is the formula for Compound ratio of a:b and c:d is ac:bd?",
        "input": "",
        "output": "If A:B = 2:3 and B:C = 4:5, then A:B:C = 8:12:15",
        "text": "### Instruction:\nWhat is the formula for Compound ratio of a:b and c:d is ac:bd?\n\n### Response:\nIf A:B = 2:3 and B:C = 4:5, then A:B:C = 8:12:15"
    },
    {
        "instruction": "What is Stochastic Gradient Descent (SGD)?",
        "input": "",
        "output": "Stochastic Gradient Descent (SGD): one sample at a time, noisy but fast",
        "text": "### Instruction:\nWhat is Stochastic Gradient Descent (SGD)?\n\n### Response:\nStochastic Gradient Descent (SGD): one sample at a time, noisy but fast"
    },
    {
        "instruction": "What is 15% of 840?",
        "output": "15% of 840 = (15/100) x 840 = 0.15 x 840 = 126",
        "text": "### Instruction:\nWhat is 15% of 840?\n\n### Response:\n15% of 840 = (15/100) x 840 = 0.15 x 840 = 126"
    },
    {
        "instruction": "What are the key concepts in Para Jumbles?",
        "input": "",
        "output": "Find opening sentence: introduces topic, no pronoun reference without antecedent\nFind closing sentence: concludes or summarizes\nLink sentences using: pronouns, connectors, time sequence\nGroup related sentences first",
        "text": "### Instruction:\nWhat are the key concepts in Para Jumbles?\n\n### Response:\nFind opening sentence: introduces topic, no pronoun reference without antecedent\nFind closing sentence: concludes or summarizes\nLink sentences using: pronouns, connectors, time sequence\nGroup related sentences first"
    },
    {
        "instruction": "What is WPI (Wholesale Price Index)?",
        "input": "",
        "output": "WPI (Wholesale Price Index): measures wholesale prices",
        "text": "### Instruction:\nWhat is WPI (Wholesale Price Index)?\n\n### Response:\nWPI (Wholesale Price Index): measures wholesale prices"
    },
    {
        "instruction": "What is Articles?",
        "input": "",
        "output": "Articles: 'a' before consonant sound, 'an' before vowel sound",
        "text": "### Instruction:\nWhat is Articles?\n\n### Response:\nArticles: 'a' before consonant sound, 'an' before vowel sound"
    },
    {
        "instruction": "If APPLE is coded as 12345, what is the code for PLEA?",
        "output": "A=1, P=2, P=2, L=3, E=4. So PLEA = P(2), L(3), E(4), A(1) = 2341",
        "text": "### Instruction:\nIf APPLE is coded as 12345, what is the code for PLEA?\n\n### Response:\nA=1, P=2, P=2, L=3, E=4. So PLEA = P(2), L(3), E(4), A(1) = 2341"
    },
    {
        "instruction": "What is IRDA/IRDAI?",
        "input": "",
        "output": "IRDA/IRDAI: Insurance Regulatory and Development Authority of India",
        "text": "### Instruction:\nWhat is IRDA/IRDAI?\n\n### Response:\nIRDA/IRDAI: Insurance Regulatory and Development Authority of India"
    },
    {
        "instruction": "What is the formula for Parth Bijpuriya | 2025?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for Parth Bijpuriya | 2025?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is AIIB?",
        "input": "",
        "output": "AIIB: Asian Infrastructure Investment Bank",
        "text": "### Instruction:\nWhat is AIIB?\n\n### Response:\nAIIB: Asian Infrastructure Investment Bank"
    },
    {
        "instruction": "What is Vocabulary in context?",
        "input": "",
        "output": "Vocabulary in context: use surrounding words to find meaning",
        "text": "### Instruction:\nWhat is Vocabulary in context?\n\n### Response:\nVocabulary in context: use surrounding words to find meaning"
    },
    {
        "instruction": "What is Primary deficit?",
        "input": "",
        "output": "Primary deficit: fiscal deficit minus interest payments",
        "text": "### Instruction:\nWhat is Primary deficit?\n\n### Response:\nPrimary deficit: fiscal deficit minus interest payments"
    },
    {
        "instruction": "What is Firewall?",
        "input": "",
        "output": "Firewall: security system to filter network traffic",
        "text": "### Instruction:\nWhat is Firewall?\n\n### Response:\nFirewall: security system to filter network traffic"
    },
    {
        "instruction": "What is the formula for Percentage means per hundred. x% = x/100?",
        "input": "",
        "output": "Percentage increase = (increase / original) x 100",
        "text": "### Instruction:\nWhat is the formula for Percentage means per hundred. x% = x/100?\n\n### Response:\nPercentage increase = (increase / original) x 100"
    },
    {
        "instruction": "What is the compound interest on Rs 10000 at 10% per annum for 2 years?",
        "output": "Amount = 10000 x (1 + 10/100)^2 = 10000 x 1.21 = 12100. CI = 12100 - 10000 = Rs 2100",
        "text": "### Instruction:\nWhat is the compound interest on Rs 10000 at 10% per annum for 2 years?\n\n### Response:\nAmount = 10000 x (1 + 10/100)^2 = 10000 x 1.21 = 12100. CI = 12100 - 10000 = Rs 2100"
    },
    {
        "instruction": "What is Phishing?",
        "input": "",
        "output": "Phishing: fake emails/websites to steal credentials",
        "text": "### Instruction:\nWhat is Phishing?\n\n### Response:\nPhishing: fake emails/websites to steal credentials"
    },
    {
        "instruction": "What is the formula for Right triangle: Area = 1/2 x base x perpendicular?",
        "input": "",
        "output": "Cone: Volume = 1/3 x pi x r^2 x h, Slant height = sqrt(r^2 + h^2)",
        "text": "### Instruction:\nWhat is the formula for Right triangle: Area = 1/2 x base x perpendicular?\n\n### Response:\nCone: Volume = 1/3 x pi x r^2 x h, Slant height = sqrt(r^2 + h^2)"
    },
    {
        "instruction": "What is the formula for --- Percentage ---?",
        "input": "",
        "output": "If price increases by r%, new price = original x (1 + r/100)",
        "text": "### Instruction:\nWhat is the formula for --- Percentage ---?\n\n### Response:\nIf price increases by r%, new price = original x (1 + r/100)"
    },
    {
        "instruction": "What is Pandas?",
        "input": "",
        "output": "Pandas: data manipulation and analysis",
        "text": "### Instruction:\nWhat is Pandas?\n\n### Response:\nPandas: data manipulation and analysis"
    },
    {
        "instruction": "What is cv2.resize()?",
        "input": "",
        "output": "cv2.resize(): resize image",
        "text": "### Instruction:\nWhat is cv2.resize()?\n\n### Response:\ncv2.resize(): resize image"
    },
    {
        "instruction": "What is For independent events?",
        "input": "",
        "output": "For independent events: P(A and B) = P(A) x P(B)",
        "text": "### Instruction:\nWhat is For independent events?\n\n### Response:\nFor independent events: P(A and B) = P(A) x P(B)"
    },
    {
        "instruction": "What is the formula for Cone: Volume = 1/3 x pi x r^2 x h, Slant height = sqrt(r^2 + h^2)?",
        "input": "",
        "output": "Cube: Volume = a^3, Surface Area = 6a^2",
        "text": "### Instruction:\nWhat is the formula for Cone: Volume = 1/3 x pi x r^2 x h, Slant height = sqrt(r^2 + h^2)?\n\n### Response:\nCube: Volume = a^3, Surface Area = 6a^2"
    },
    {
        "instruction": "What are the key concepts in Monetary Policy Tools?",
        "input": "",
        "output": "Repo Rate: rate at which RBI lends to commercial banks\n  - Increase in repo rate = credit becomes expensive = inflation control\nReverse Repo Rate: rate at which RBI borrows from banks\n  - Always lower than repo rate\nCash Reserve Ratio (CRR): percentage of deposits banks must keep with RBI\n  - No interest earned on CRR\n  - Increase in CRR = less money with banks = less credit = inflation control\nStatutory Liquidity Ratio (SLR): percentage of deposits banks must maintain\n  in liquid assets (gold, govt securities)\n  - Maximum SLR can be 40%\nBank Rate: rate at which RBI provides long-term loans to",
        "text": "### Instruction:\nWhat are the key concepts in Monetary Policy Tools?\n\n### Response:\nRepo Rate: rate at which RBI lends to commercial banks\n  - Increase in repo rate = credit becomes expensive = inflation control\nReverse Repo Rate: rate at which RBI borrows from banks\n  - Always lower than repo rate\nCash Reserve Ratio (CRR): percentage of deposits banks must keep with RBI\n  - No interest earned on CRR\n  - Increase in CRR = less money with banks = less credit = inflation control\nStatutory Liquidity Ratio (SLR): percentage of deposits banks must maintain\n  in liquid assets (gold, govt securities)\n  - Maximum SLR can be 40%\nBank Rate: rate at which RBI provides long-term loans to"
    },
    {
        "instruction": "What is Leaky ReLU?",
        "input": "",
        "output": "Leaky ReLU: allows small negative values",
        "text": "### Instruction:\nWhat is Leaky ReLU?\n\n### Response:\nLeaky ReLU: allows small negative values"
    },
    {
        "instruction": "What is Cheque?",
        "input": "",
        "output": "Cheque: written order to bank to pay amount from account",
        "text": "### Instruction:\nWhat is Cheque?\n\n### Response:\nCheque: written order to bank to pay amount from account"
    },
    {
        "instruction": "What is the formula for SECTION 2: REASONING ABILITY?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for SECTION 2: REASONING ABILITY?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is Open Market Operations (OMO)?",
        "input": "",
        "output": "Open Market Operations (OMO): RBI buys/sells govt securities to control liquidity",
        "text": "### Instruction:\nWhat is Open Market Operations (OMO)?\n\n### Response:\nOpen Market Operations (OMO): RBI buys/sells govt securities to control liquidity"
    },
    {
        "instruction": "What are the key concepts in Machine Learning Fundamentals?",
        "input": "",
        "output": "Machine Learning: systems that learn from data without explicit programming\nTypes of ML:\n  1. Supervised Learning: labeled data, predict output (classification, regression)\n  2. Unsupervised Learning: unlabeled data, find patterns (clustering, dimensionality reduction)\n  3. Reinforcement Learning: agent learns by reward/punishment\n  4. Semi-supervised: mix of labeled and unlabeled data\n\nOverfitting: model performs well on training data but poorly on new data\n  - Solution: more data, regularization, dropout, early stopping\nUnderfitting: model too simple, performs poorly on both train and test",
        "text": "### Instruction:\nWhat are the key concepts in Machine Learning Fundamentals?\n\n### Response:\nMachine Learning: systems that learn from data without explicit programming\nTypes of ML:\n  1. Supervised Learning: labeled data, predict output (classification, regression)\n  2. Unsupervised Learning: unlabeled data, find patterns (clustering, dimensionality reduction)\n  3. Reinforcement Learning: agent learns by reward/punishment\n  4. Semi-supervised: mix of labeled and unlabeled data\n\nOverfitting: model performs well on training data but poorly on new data\n  - Solution: more data, regularization, dropout, early stopping\nUnderfitting: model too simple, performs poorly on both train and test"
    },
    {
        "instruction": "What are the key concepts in NeuroSpeech (AI Speech Therapy)?",
        "input": "",
        "output": "What: AI-powered speech therapy platform for people with speech disorders\nTech: Python, Streamlit, Supabase, OpenAI Whisper, HuggingFace Spaces\nKey Features:\n  - Real-time speech-to-text using Whisper (transformer-based ASR)\n  - IPA phonetic analysis for pronunciation scoring\n  - Supabase Auth with email and Google OAuth\n  - PostgreSQL with Row Level Security (users see only their data)\n  - Realtime subscriptions for live leaderboard updates\n  - Edge Functions (Deno) for AI feedback and progress reports\n  - Storage for audio recordings\n  - Streak tracking with PostgreSQL functions\n  - Admin pa",
        "text": "### Instruction:\nWhat are the key concepts in NeuroSpeech (AI Speech Therapy)?\n\n### Response:\nWhat: AI-powered speech therapy platform for people with speech disorders\nTech: Python, Streamlit, Supabase, OpenAI Whisper, HuggingFace Spaces\nKey Features:\n  - Real-time speech-to-text using Whisper (transformer-based ASR)\n  - IPA phonetic analysis for pronunciation scoring\n  - Supabase Auth with email and Google OAuth\n  - PostgreSQL with Row Level Security (users see only their data)\n  - Realtime subscriptions for live leaderboard updates\n  - Edge Functions (Deno) for AI feedback and progress reports\n  - Storage for audio recordings\n  - Streak tracking with PostgreSQL functions\n  - Admin pa"
    },
    {
        "instruction": "What is Major rivers?",
        "input": "",
        "output": "Major rivers: Ganga, Yamuna, Brahmaputra, Godavari, Krishna, Kaveri",
        "text": "### Instruction:\nWhat is Major rivers?\n\n### Response:\nMajor rivers: Ganga, Yamuna, Brahmaputra, Godavari, Krishna, Kaveri"
    },
    {
        "instruction": "What is Thar Desert?",
        "input": "",
        "output": "Thar Desert: in Rajasthan, only hot desert in India",
        "text": "### Instruction:\nWhat is Thar Desert?\n\n### Response:\nThar Desert: in Rajasthan, only hot desert in India"
    },
    {
        "instruction": "What are the key concepts in Important Banking Terms?",
        "input": "",
        "output": "NPA (Non-Performing Asset): loan where interest/principal overdue 90+ days\nSubstandard Asset: NPA for less than 12 months\nDoubtful Asset: NPA for more than 12 months\nLoss Asset: uncollectible NPA\nSARFAESI Act: allows banks to recover NPA without court intervention\nCIBIL Score: credit score ranging 300-900 (750+ considered good)\nKYC: Know Your Customer - identity verification process\nNEFT: National Electronic Funds Transfer (settled in batches)\nRTGS: Real Time Gross Settlement (minimum 2 lakh, real time)\nIMPS: Immediate Payment Service (24x7, instant, up to 5 lakh)\nUPI: Unified Payments Interfa",
        "text": "### Instruction:\nWhat are the key concepts in Important Banking Terms?\n\n### Response:\nNPA (Non-Performing Asset): loan where interest/principal overdue 90+ days\nSubstandard Asset: NPA for less than 12 months\nDoubtful Asset: NPA for more than 12 months\nLoss Asset: uncollectible NPA\nSARFAESI Act: allows banks to recover NPA without court intervention\nCIBIL Score: credit score ranging 300-900 (750+ considered good)\nKYC: Know Your Customer - identity verification process\nNEFT: National Electronic Funds Transfer (settled in batches)\nRTGS: Real Time Gross Settlement (minimum 2 lakh, real time)\nIMPS: Immediate Payment Service (24x7, instant, up to 5 lakh)\nUPI: Unified Payments Interfa"
    },
    {
        "instruction": "What is the formula for --- Ratio and Proportion ---?",
        "input": "",
        "output": "If a:b = c:d then ad = bc (cross multiplication)",
        "text": "### Instruction:\nWhat is the formula for --- Ratio and Proportion ---?\n\n### Response:\nIf a:b = c:d then ad = bc (cross multiplication)"
    },
    {
        "instruction": "What is Symbol coding?",
        "input": "",
        "output": "Symbol coding: letters replaced by symbols",
        "text": "### Instruction:\nWhat is Symbol coding?\n\n### Response:\nSymbol coding: letters replaced by symbols"
    },
    {
        "instruction": "What is Foreign banks?",
        "input": "",
        "output": "Foreign banks: headquartered abroad (Citibank, HSBC)",
        "text": "### Instruction:\nWhat is Foreign banks?\n\n### Response:\nForeign banks: headquartered abroad (Citibank, HSBC)"
    },
    {
        "instruction": "What is NHB?",
        "input": "",
        "output": "NHB: National Housing Bank - regulates housing finance companies",
        "text": "### Instruction:\nWhat is NHB?\n\n### Response:\nNHB: National Housing Bank - regulates housing finance companies"
    },
    {
        "instruction": "What is Based on Bayes theorem?",
        "input": "",
        "output": "Based on Bayes theorem: P(A|B) = P(B|A) x P(A) / P(B)",
        "text": "### Instruction:\nWhat is Based on Bayes theorem?\n\n### Response:\nBased on Bayes theorem: P(A|B) = P(B|A) x P(A) / P(B)"
    },
    {
        "instruction": "What is the difference between LSTM and Transformer?",
        "input": "",
        "output": "LSTM processes sequences step by step, has memory through cell state,\nstruggles with very long dependencies due to vanishing gradients.\nTransformer processes all tokens in parallel using self-attention,\nhandles long dependencies easily, much faster to train, but needs more data.\nTransformers have replaced LSTMs in most NLP tasks.\n\n================================================================================\nSECTION 8: PARTH'S PROJECT SUMMARIES (FOR INTERVIEWS)\n================================================================================\n\n--- NeuroSpeech (AI Speech Therapy) ---\nWhat: AI-powered speech therapy platform for people with speech disorders\nTech: Python, Streamlit, Supabase, OpenAI Whisper, HuggingFace Spaces\nKey Features:\n  - Real-time speech-to-text using Whisper (transformer-based ASR)\n  - IPA phonetic analysis for pronunciation scoring\n  - Supabase Auth with email and Google OAuth\n  - PostgreSQL with Row Level Security (users see only their data)\n  - Realtime subscriptions for live leaderboard updates\n  - Edge Functions (Deno) for AI feedback and progress reports\n  - Storage for audio recordings\n  - Streak tracking with PostgreSQL functions\n  - Admin panel with service role\nLive: https://huggingface.co/spaces/parthbijpuriya/neurospeech\nImpact: Accessible speech therapy without expensive therapists\n\n--- Drive-Awake (Drowsiness Detection) ---\nWhat: Real-time driver drowsiness detection system\nTech: Python, TensorFlow, FastAPI, OpenCV, pygame, Docker\nKey Features:\n  - CNN model for eye state classification (open/closed)\n  - Haar Cascade for face detection\n  - Eye Aspect Ratio (EAR) for blink pattern analysis\n  - Audio alarm when drowsiness detected\n  - 92%+ accuracy, real-time at 30+ FPS\n  - FastAPI for REST API deployment\n  - Docker containerization\nImpact: Prevents road accidents from driver fatigue\nBugs to fix: duplicate requirements, Dockerfile OS mismatch, float() on array\n\n--- Emotion Detector ---\nWhat: Real-time emotion detection with recommendations\nTech: TensorFlow, OpenCV, PyQt5\nKey Features:\n  - CNN for facial emotion recognition (7 emotions)\n  - PyQt5 desktop GUI\n  - Emotion-based content recommender (100+ suggestions)\n  - Sub-second inference\n\n--- Music Recommender ---\nWhat: Personalized music recommendation system\nTech: Flask, Spotify API, scikit-learn\nKey Features:\n  - K-means clustering for music grouping\n  - Collaborative filtering for personalization\n  - Spotify API for music data\n  - 28% accuracy improvement over baseline\n  - Tested with 1000+ users\n  - Under 2 second inference\n\n================================================================================\nEND OF NOTES\n================================================================================",
        "text": "### Instruction:\nWhat is the difference between LSTM and Transformer?\n\n### Response:\nLSTM processes sequences step by step, has memory through cell state,\nstruggles with very long dependencies due to vanishing gradients.\nTransformer processes all tokens in parallel using self-attention,\nhandles long dependencies easily, much faster to train, but needs more data.\nTransformers have replaced LSTMs in most NLP tasks.\n\n================================================================================\nSECTION 8: PARTH'S PROJECT SUMMARIES (FOR INTERVIEWS)\n================================================================================\n\n--- NeuroSpeech (AI Speech Therapy) ---\nWhat: AI-powered speech therapy platform for people with speech disorders\nTech: Python, Streamlit, Supabase, OpenAI Whisper, HuggingFace Spaces\nKey Features:\n  - Real-time speech-to-text using Whisper (transformer-based ASR)\n  - IPA phonetic analysis for pronunciation scoring\n  - Supabase Auth with email and Google OAuth\n  - PostgreSQL with Row Level Security (users see only their data)\n  - Realtime subscriptions for live leaderboard updates\n  - Edge Functions (Deno) for AI feedback and progress reports\n  - Storage for audio recordings\n  - Streak tracking with PostgreSQL functions\n  - Admin panel with service role\nLive: https://huggingface.co/spaces/parthbijpuriya/neurospeech\nImpact: Accessible speech therapy without expensive therapists\n\n--- Drive-Awake (Drowsiness Detection) ---\nWhat: Real-time driver drowsiness detection system\nTech: Python, TensorFlow, FastAPI, OpenCV, pygame, Docker\nKey Features:\n  - CNN model for eye state classification (open/closed)\n  - Haar Cascade for face detection\n  - Eye Aspect Ratio (EAR) for blink pattern analysis\n  - Audio alarm when drowsiness detected\n  - 92%+ accuracy, real-time at 30+ FPS\n  - FastAPI for REST API deployment\n  - Docker containerization\nImpact: Prevents road accidents from driver fatigue\nBugs to fix: duplicate requirements, Dockerfile OS mismatch, float() on array\n\n--- Emotion Detector ---\nWhat: Real-time emotion detection with recommendations\nTech: TensorFlow, OpenCV, PyQt5\nKey Features:\n  - CNN for facial emotion recognition (7 emotions)\n  - PyQt5 desktop GUI\n  - Emotion-based content recommender (100+ suggestions)\n  - Sub-second inference\n\n--- Music Recommender ---\nWhat: Personalized music recommendation system\nTech: Flask, Spotify API, scikit-learn\nKey Features:\n  - K-means clustering for music grouping\n  - Collaborative filtering for personalization\n  - Spotify API for music data\n  - 28% accuracy improvement over baseline\n  - Tested with 1000+ users\n  - Under 2 second inference\n\n================================================================================\nEND OF NOTES\n================================================================================"
    },
    {
        "instruction": "What is Number coding?",
        "input": "",
        "output": "Number coding: words replaced by numbers",
        "text": "### Instruction:\nWhat is Number coding?\n\n### Response:\nNumber coding: words replaced by numbers"
    },
    {
        "instruction": "What is Sentiment Analysis?",
        "input": "",
        "output": "Sentiment Analysis: classify text as positive/negative/neutral",
        "text": "### Instruction:\nWhat is Sentiment Analysis?\n\n### Response:\nSentiment Analysis: classify text as positive/negative/neutral"
    },
    {
        "instruction": "What is PyTorch?",
        "input": "",
        "output": "PyTorch: deep learning framework, dynamic computation graphs",
        "text": "### Instruction:\nWhat is PyTorch?\n\n### Response:\nPyTorch: deep learning framework, dynamic computation graphs"
    },
    {
        "instruction": "If a man walks at 5 km/h and a woman walks at 3 km/h in opposite directions, what is their relative speed?",
        "output": "Relative speed (opposite directions) = 5 + 3 = 8 km/h",
        "text": "### Instruction:\nIf a man walks at 5 km/h and a woman walks at 3 km/h in opposite directions, what is their relative speed?\n\n### Response:\nRelative speed (opposite directions) = 5 + 3 = 8 km/h"
    },
    {
        "instruction": "What is PFRDA?",
        "input": "",
        "output": "PFRDA: Pension Fund Regulatory and Development Authority",
        "text": "### Instruction:\nWhat is PFRDA?\n\n### Response:\nPFRDA: Pension Fund Regulatory and Development Authority"
    },
    {
        "instruction": "What is RTGS?",
        "input": "",
        "output": "RTGS: Real Time Gross Settlement (minimum 2 lakh, real time)",
        "text": "### Instruction:\nWhat is RTGS?\n\n### Response:\nRTGS: Real Time Gross Settlement (minimum 2 lakh, real time)"
    },
    {
        "instruction": "What is Fiscal deficit?",
        "input": "",
        "output": "Fiscal deficit: government expenditure exceeds revenue",
        "text": "### Instruction:\nWhat is Fiscal deficit?\n\n### Response:\nFiscal deficit: government expenditure exceeds revenue"
    },
    {
        "instruction": "What is Fibonacci?",
        "input": "",
        "output": "Fibonacci: each term = sum of previous two (1,1,2,3,5,8...)",
        "text": "### Instruction:\nWhat is Fibonacci?\n\n### Response:\nFibonacci: each term = sum of previous two (1,1,2,3,5,8...)"
    },
    {
        "instruction": "What is GNP?",
        "input": "",
        "output": "GNP: Gross National Product - GDP + income from abroad - income to foreigners",
        "text": "### Instruction:\nWhat is GNP?\n\n### Response:\nGNP: Gross National Product - GDP + income from abroad - income to foreigners"
    },
    {
        "instruction": "What is Inertia (within-cluster sum of squares)?",
        "input": "",
        "output": "Inertia (within-cluster sum of squares): lower is better",
        "text": "### Instruction:\nWhat is Inertia (within-cluster sum of squares)?\n\n### Response:\nInertia (within-cluster sum of squares): lower is better"
    },
    {
        "instruction": "What is Inception/GoogLeNet?",
        "input": "",
        "output": "Inception/GoogLeNet: parallel convolutions",
        "text": "### Instruction:\nWhat is Inception/GoogLeNet?\n\n### Response:\nInception/GoogLeNet: parallel convolutions"
    },
    {
        "instruction": "What is Marked Price?",
        "input": "",
        "output": "Marked Price: price written on item before discount",
        "text": "### Instruction:\nWhat is Marked Price?\n\n### Response:\nMarked Price: price written on item before discount"
    },
    {
        "instruction": "What is Drive-Awake?",
        "input": "",
        "output": "Drive-Awake: CNN for eye state classification (open/closed)",
        "text": "### Instruction:\nWhat is Drive-Awake?\n\n### Response:\nDrive-Awake: CNN for eye state classification (open/closed)"
    },
    {
        "instruction": "What is the formula for --- Probability ---?",
        "input": "",
        "output": "P(A) + P(not A) = 1",
        "text": "### Instruction:\nWhat is the formula for --- Probability ---?\n\n### Response:\nP(A) + P(not A) = 1"
    },
    {
        "instruction": "What is Trojan?",
        "input": "",
        "output": "Trojan: disguises as legitimate software",
        "text": "### Instruction:\nWhat is Trojan?\n\n### Response:\nTrojan: disguises as legitimate software"
    },
    {
        "instruction": "What is Moving same direction?",
        "input": "",
        "output": "Moving same direction: Time = (L1+L2) / (S1-S2)",
        "text": "### Instruction:\nWhat is Moving same direction?\n\n### Response:\nMoving same direction: Time = (L1+L2) / (S1-S2)"
    },
    {
        "instruction": "What are the key concepts in Indian Economy?",
        "input": "",
        "output": "GDP: Gross Domestic Product - total value of goods/services in a year\nGNP: Gross National Product - GDP + income from abroad - income to foreigners\nNNP: Net National Product = GNP - Depreciation\nPer capita income: NNP / Population\nFiscal deficit: government expenditure exceeds revenue\nRevenue deficit: revenue expenditure exceeds revenue receipts\nPrimary deficit: fiscal deficit minus interest payments\nNITI Aayog: replaced Planning Commission in 2015\nFive Year Plans: replaced by 3-year action plan + 7-year strategy\nIndia's GDP growth rate: around 6-7% annually\nIndia's inflation measure: CPI (Con",
        "text": "### Instruction:\nWhat are the key concepts in Indian Economy?\n\n### Response:\nGDP: Gross Domestic Product - total value of goods/services in a year\nGNP: Gross National Product - GDP + income from abroad - income to foreigners\nNNP: Net National Product = GNP - Depreciation\nPer capita income: NNP / Population\nFiscal deficit: government expenditure exceeds revenue\nRevenue deficit: revenue expenditure exceeds revenue receipts\nPrimary deficit: fiscal deficit minus interest payments\nNITI Aayog: replaced Planning Commission in 2015\nFive Year Plans: replaced by 3-year action plan + 7-year strategy\nIndia's GDP growth rate: around 6-7% annually\nIndia's inflation measure: CPI (Con"
    },
    {
        "instruction": "What is the formula for SECTION 5: GENERAL AWARENESS / CURRENT AFFAIRS (STATIC GK)?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for SECTION 5: GENERAL AWARENESS / CURRENT AFFAIRS (STATIC GK)?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is the formula for --- Simple Interest ---?",
        "input": "",
        "output": "Amount = Principal + SI",
        "text": "### Instruction:\nWhat is the formula for --- Simple Interest ---?\n\n### Response:\nAmount = Principal + SI"
    },
    {
        "instruction": "What is the formula for Starting direction: North?",
        "input": "",
        "output": "After 180 degree turn from North = South",
        "text": "### Instruction:\nWhat is the formula for Starting direction: North?\n\n### Response:\nAfter 180 degree turn from North = South"
    },
    {
        "instruction": "What is the formula for MY NOTES \u2014 Banking PO (IBPS/SBI) + AI/ML Interview Prep?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for MY NOTES \u2014 Banking PO (IBPS/SBI) + AI/ML Interview Prep?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is the formula for Composite numbers: more than 2 factors. Example: 4, 6, 8, 9?",
        "input": "",
        "output": "HCF x LCM = Product of two numbers",
        "text": "### Instruction:\nWhat is the formula for Composite numbers: more than 2 factors. Example: 4, 6, 8, 9?\n\n### Response:\nHCF x LCM = Product of two numbers"
    },
    {
        "instruction": "What is SMTP?",
        "input": "",
        "output": "SMTP: Simple Mail Transfer Protocol - sending emails",
        "text": "### Instruction:\nWhat is SMTP?\n\n### Response:\nSMTP: Simple Mail Transfer Protocol - sending emails"
    },
    {
        "instruction": "What is Shishu?",
        "input": "",
        "output": "Shishu: up to 50,000",
        "text": "### Instruction:\nWhat is Shishu?\n\n### Response:\nShishu: up to 50,000"
    },
    {
        "instruction": "What is Decision Tree?",
        "input": "",
        "output": "Decision Tree: tree structure, splits data on features",
        "text": "### Instruction:\nWhat is Decision Tree?\n\n### Response:\nDecision Tree: tree structure, splits data on features"
    },
    {
        "instruction": "What is the formula for Triangle: Area = 1/2 x base x height?",
        "input": "",
        "output": "Cylinder: Volume = pi x r^2 x h, Curved Surface Area = 2 x pi x r x h",
        "text": "### Instruction:\nWhat is the formula for Triangle: Area = 1/2 x base x height?\n\n### Response:\nCylinder: Volume = pi x r^2 x h, Curved Surface Area = 2 x pi x r x h"
    },
    {
        "instruction": "What is Prepositions?",
        "input": "",
        "output": "Prepositions: in, on, at, by, for, with, about, above, below",
        "text": "### Instruction:\nWhat is Prepositions?\n\n### Response:\nPrepositions: in, on, at, by, for, with, about, above, below"
    },
    {
        "instruction": "What is attention mechanism in transformers?",
        "input": "",
        "output": "Attention allows each token to look at all other tokens when building\nits representation. Attention(Q,K,V) = softmax(QK^T / sqrt(d_k)) V\nQuery asks what to look for, Key says what each token offers,\nValue is the actual information. Multi-head attention runs multiple\nattention operations in parallel for different aspects.",
        "text": "### Instruction:\nWhat is attention mechanism in transformers?\n\n### Response:\nAttention allows each token to look at all other tokens when building\nits representation. Attention(Q,K,V) = softmax(QK^T / sqrt(d_k)) V\nQuery asks what to look for, Key says what each token offers,\nValue is the actual information. Multi-head attention runs multiple\nattention operations in parallel for different aspects."
    },
    {
        "instruction": "What is the formula for Probability = Favourable outcomes / Total outcomes?",
        "input": "",
        "output": "P(A or B) = P(A) + P(B) - P(A and B)",
        "text": "### Instruction:\nWhat is the formula for Probability = Favourable outcomes / Total outcomes?\n\n### Response:\nP(A or B) = P(A) + P(B) - P(A and B)"
    },
    {
        "instruction": "What is the significance/value of Article 22?",
        "input": "",
        "output": "Article 22: Protection against arbitrary arrest",
        "text": "### Instruction:\nWhat is the significance/value of Article 22?\n\n### Response:\nArticle 22: Protection against arbitrary arrest"
    },
    {
        "instruction": "What is the formula for Rectangle: Area = length x breadth, Perimeter = 2(l+b)?",
        "input": "",
        "output": "Right triangle: Area = 1/2 x base x perpendicular",
        "text": "### Instruction:\nWhat is the formula for Rectangle: Area = length x breadth, Perimeter = 2(l+b)?\n\n### Response:\nRight triangle: Area = 1/2 x base x perpendicular"
    },
    {
        "instruction": "What are the key concepts in Basics?",
        "input": "",
        "output": "CPU: Central Processing Unit - brain of computer\nRAM: Random Access Memory - temporary storage, volatile\nROM: Read Only Memory - permanent storage, non-volatile\nHard Disk: permanent storage\nCache: fastest memory, stores frequently used data\nRegister: fastest, inside CPU\nStorage hierarchy: Register > Cache > RAM > Hard Disk",
        "text": "### Instruction:\nWhat are the key concepts in Basics?\n\n### Response:\nCPU: Central Processing Unit - brain of computer\nRAM: Random Access Memory - temporary storage, volatile\nROM: Read Only Memory - permanent storage, non-volatile\nHard Disk: permanent storage\nCache: fastest memory, stores frequently used data\nRegister: fastest, inside CPU\nStorage hierarchy: Register > Cache > RAM > Hard Disk"
    },
    {
        "instruction": "What is Object Detection?",
        "input": "",
        "output": "Object Detection: locate and classify objects in image",
        "text": "### Instruction:\nWhat is Object Detection?\n\n### Response:\nObject Detection: locate and classify objects in image"
    },
    {
        "instruction": "What is If A is x% more than B, B is less than A by?",
        "input": "",
        "output": "If A is x% more than B, B is less than A by: (x / 100+x) x 100 percent",
        "text": "### Instruction:\nWhat is If A is x% more than B, B is less than A by?\n\n### Response:\nIf A is x% more than B, B is less than A by: (x / 100+x) x 100 percent"
    },
    {
        "instruction": "What are the key concepts in CNN (Convolutional Neural Networks)?",
        "input": "",
        "output": "Used for: image classification, object detection, computer vision\nConvolutional Layer: applies filters to detect features (edges, textures)\nPooling Layer: reduces spatial dimensions (Max Pooling, Average Pooling)\nFully Connected Layer: final classification layer\nFeature Maps: output of convolutional layer\nStride: step size of filter movement\nPadding: adds zeros around input to preserve spatial dimensions\nSame padding: output same size as input\nValid padding: no padding, output smaller\n\nFamous CNNs:\n  - LeNet: first CNN (1998)\n  - AlexNet: won ImageNet 2012, made deep learning popular\n  - VGG:",
        "text": "### Instruction:\nWhat are the key concepts in CNN (Convolutional Neural Networks)?\n\n### Response:\nUsed for: image classification, object detection, computer vision\nConvolutional Layer: applies filters to detect features (edges, textures)\nPooling Layer: reduces spatial dimensions (Max Pooling, Average Pooling)\nFully Connected Layer: final classification layer\nFeature Maps: output of convolutional layer\nStride: step size of filter movement\nPadding: adds zeros around input to preserve spatial dimensions\nSame padding: output same size as input\nValid padding: no padding, output smaller\n\nFamous CNNs:\n  - LeNet: first CNN (1998)\n  - AlexNet: won ImageNet 2012, made deep learning popular\n  - VGG:"
    },
    {
        "instruction": "What is Article 23?",
        "input": "",
        "output": "Article 23: Prohibition of traffic in human beings",
        "text": "### Instruction:\nWhat is Article 23?\n\n### Response:\nArticle 23: Prohibition of traffic in human beings"
    },
    {
        "instruction": "What is Sigmoid?",
        "input": "",
        "output": "Sigmoid: 1/(1+e^-x), output 0 to 1, vanishing gradient problem",
        "text": "### Instruction:\nWhat is Sigmoid?\n\n### Response:\nSigmoid: 1/(1+e^-x), output 0 to 1, vanishing gradient problem"
    },
    {
        "instruction": "What is the formula for Cube: Volume = a^3, Surface Area = 6a^2?",
        "input": "",
        "output": "Probability = Favourable outcomes / Total outcomes",
        "text": "### Instruction:\nWhat is the formula for Cube: Volume = a^3, Surface Area = 6a^2?\n\n### Response:\nProbability = Favourable outcomes / Total outcomes"
    },
    {
        "instruction": "What are the key concepts in Averages?",
        "input": "",
        "output": "Average = Sum of observations / Number of observations\nSum = Average x Number\nIf average of n numbers is A and one number x is replaced by y:\n  New average = A + (y-x)/n\nWeighted average = (w1*x1 + w2*x2 + ...) / (w1 + w2 + ...)",
        "text": "### Instruction:\nWhat are the key concepts in Averages?\n\n### Response:\nAverage = Sum of observations / Number of observations\nSum = Average x Number\nIf average of n numbers is A and one number x is replaced by y:\n  New average = A + (y-x)/n\nWeighted average = (w1*x1 + w2*x2 + ...) / (w1 + w2 + ...)"
    },
    {
        "instruction": "What is Divisibility by 9?",
        "input": "",
        "output": "Divisibility by 9: sum of digits divisible by 9",
        "text": "### Instruction:\nWhat is Divisibility by 9?\n\n### Response:\nDivisibility by 9: sum of digits divisible by 9"
    },
    {
        "instruction": "What is Used in?",
        "input": "",
        "output": "Used in: Music Recommender project (Parth's project)",
        "text": "### Instruction:\nWhat is Used in?\n\n### Response:\nUsed in: Music Recommender project (Parth's project)"
    },
    {
        "instruction": "What is the formula for --- Blood Relations ---?",
        "input": "",
        "output": "Mother's father = Maternal grandfather",
        "text": "### Instruction:\nWhat is the formula for --- Blood Relations ---?\n\n### Response:\nMother's father = Maternal grandfather"
    },
    {
        "instruction": "What is the formula for Mean proportion of a and b = sqrt(ab)?",
        "input": "",
        "output": "Average = Sum of observations / Number of observations",
        "text": "### Instruction:\nWhat is the formula for Mean proportion of a and b = sqrt(ab)?\n\n### Response:\nAverage = Sum of observations / Number of observations"
    },
    {
        "instruction": "What is Inference questions?",
        "input": "",
        "output": "Inference questions: answer not directly stated, must be logically drawn",
        "text": "### Instruction:\nWhat is Inference questions?\n\n### Response:\nInference questions: answer not directly stated, must be logically drawn"
    },
    {
        "instruction": "What are the key concepts in Grammar Rules?",
        "input": "",
        "output": "Subject-verb agreement: singular subject takes singular verb\nArticles: 'a' before consonant sound, 'an' before vowel sound\nTenses: present, past, future (simple, continuous, perfect, perfect continuous)\nActive voice: subject does action. Passive voice: subject receives action\nDirect speech: exact words in quotes. Indirect: reported speech\nConjunctions: and, but, or, nor, for, yet, so (FANBOYS)\nPrepositions: in, on, at, by, for, with, about, above, below",
        "text": "### Instruction:\nWhat are the key concepts in Grammar Rules?\n\n### Response:\nSubject-verb agreement: singular subject takes singular verb\nArticles: 'a' before consonant sound, 'an' before vowel sound\nTenses: present, past, future (simple, continuous, perfect, perfect continuous)\nActive voice: subject does action. Passive voice: subject receives action\nDirect speech: exact words in quotes. Indirect: reported speech\nConjunctions: and, but, or, nor, for, yet, so (FANBOYS)\nPrepositions: in, on, at, by, for, with, about, above, below"
    },
    {
        "instruction": "What is the formula for Discount% = (Discount / Marked Price) x 100?",
        "input": "",
        "output": "Simple Interest (SI) = (Principal x Rate x Time) / 100",
        "text": "### Instruction:\nWhat is the formula for Discount% = (Discount / Marked Price) x 100?\n\n### Response:\nSimple Interest (SI) = (Principal x Rate x Time) / 100"
    },
    {
        "instruction": "What is the formula for A = B means equal?",
        "input": "",
        "output": "If A > B and B >= C, then A > C is definitely true",
        "text": "### Instruction:\nWhat is the formula for A = B means equal?\n\n### Response:\nIf A > B and B >= C, then A > C is definitely true"
    },
    {
        "instruction": "What is 5 September?",
        "input": "",
        "output": "5 September: Teachers Day (Radhakrishnan's birthday)",
        "text": "### Instruction:\nWhat is 5 September?\n\n### Response:\n5 September: Teachers Day (Radhakrishnan's birthday)"
    },
    {
        "instruction": "What is Commercial Banks?",
        "input": "",
        "output": "Commercial Banks: SBI, PNB, HDFC, ICICI, Axis",
        "text": "### Instruction:\nWhat is Commercial Banks?\n\n### Response:\nCommercial Banks: SBI, PNB, HDFC, ICICI, Axis"
    },
    {
        "instruction": "What is NeuroSpeech?",
        "input": "",
        "output": "NeuroSpeech: OpenAI Whisper for speech-to-text (transformer-based)",
        "text": "### Instruction:\nWhat is NeuroSpeech?\n\n### Response:\nNeuroSpeech: OpenAI Whisper for speech-to-text (transformer-based)"
    },
    {
        "instruction": "What is Sphere?",
        "input": "",
        "output": "Sphere: Volume = 4/3 x pi x r^3, Surface Area = 4 x pi x r^2",
        "text": "### Instruction:\nWhat is Sphere?\n\n### Response:\nSphere: Volume = 4/3 x pi x r^3, Surface Area = 4 x pi x r^2"
    },
    {
        "instruction": "What are the key concepts in Networking?",
        "input": "",
        "output": "LAN: Local Area Network - small area (office, building)\nMAN: Metropolitan Area Network - city level\nWAN: Wide Area Network - country/world level (Internet is WAN)\nHTTP: HyperText Transfer Protocol - web browsing\nHTTPS: Secure version of HTTP\nFTP: File Transfer Protocol\nSMTP: Simple Mail Transfer Protocol - sending emails\nPOP3/IMAP: receiving emails\nIP Address: unique address for every device on internet\nIPv4: 32-bit address (4 groups of numbers, e.g., 192.168.1.1)\nIPv6: 128-bit address (solves IPv4 exhaustion)\nDNS: Domain Name System - converts domain names to IP addresses\nFirewall: security s",
        "text": "### Instruction:\nWhat are the key concepts in Networking?\n\n### Response:\nLAN: Local Area Network - small area (office, building)\nMAN: Metropolitan Area Network - city level\nWAN: Wide Area Network - country/world level (Internet is WAN)\nHTTP: HyperText Transfer Protocol - web browsing\nHTTPS: Secure version of HTTP\nFTP: File Transfer Protocol\nSMTP: Simple Mail Transfer Protocol - sending emails\nPOP3/IMAP: receiving emails\nIP Address: unique address for every device on internet\nIPv4: 32-bit address (4 groups of numbers, e.g., 192.168.1.1)\nIPv6: 128-bit address (solves IPv4 exhaustion)\nDNS: Domain Name System - converts domain names to IP addresses\nFirewall: security s"
    },
    {
        "instruction": "What is Natural numbers?",
        "input": "",
        "output": "Natural numbers: 1, 2, 3, 4, ...",
        "text": "### Instruction:\nWhat is Natural numbers?\n\n### Response:\nNatural numbers: 1, 2, 3, 4, ..."
    },
    {
        "instruction": "What is SARFAESI Act?",
        "input": "",
        "output": "SARFAESI Act: allows banks to recover NPA without court intervention",
        "text": "### Instruction:\nWhat is SARFAESI Act?\n\n### Response:\nSARFAESI Act: allows banks to recover NPA without court intervention"
    },
    {
        "instruction": "A shopkeeper buys an item for Rs 200 and sells it for Rs 250. What is the profit percentage?",
        "output": "Profit = 250 - 200 = 50. Profit% = (50/200) x 100 = 25%",
        "text": "### Instruction:\nA shopkeeper buys an item for Rs 200 and sells it for Rs 250. What is the profit percentage?\n\n### Response:\nProfit = 250 - 200 = 50. Profit% = (50/200) x 100 = 25%"
    },
    {
        "instruction": "What is Atal Pension Yojana (APY)?",
        "input": "",
        "output": "Atal Pension Yojana (APY): pension scheme for unorganised sector",
        "text": "### Instruction:\nWhat is Atal Pension Yojana (APY)?\n\n### Response:\nAtal Pension Yojana (APY): pension scheme for unorganised sector"
    },
    {
        "instruction": "What is Article 352?",
        "input": "",
        "output": "Article 352: National Emergency (armed rebellion or external aggression)",
        "text": "### Instruction:\nWhat is Article 352?\n\n### Response:\nArticle 352: National Emergency (armed rebellion or external aggression)"
    },
    {
        "instruction": "What is the formula for - Accuracy = (TP + TN) / (TP + TN + FP + FN)?",
        "input": "",
        "output": "- F1 Score = 2 x (Precision x Recall) / (Precision + Recall)",
        "text": "### Instruction:\nWhat is the formula for - Accuracy = (TP + TN) / (TP + TN + FP + FN)?\n\n### Response:\n- F1 Score = 2 x (Precision x Recall) / (Precision + Recall)"
    },
    {
        "instruction": "What is the formula for Divisibility by 11: difference of alternate digit sums divisible by 11?",
        "input": "",
        "output": "Percentage means per hundred. x% = x/100",
        "text": "### Instruction:\nWhat is the formula for Divisibility by 11: difference of alternate digit sums divisible by 11?\n\n### Response:\nPercentage means per hundred. x% = x/100"
    },
    {
        "instruction": "What is HuggingFace?",
        "input": "",
        "output": "HuggingFace: platform for sharing ML models and spaces",
        "text": "### Instruction:\nWhat is HuggingFace?\n\n### Response:\nHuggingFace: platform for sharing ML models and spaces"
    },
    {
        "instruction": "What is Underfitting?",
        "input": "",
        "output": "Underfitting: model too simple, performs poorly on both train and test",
        "text": "### Instruction:\nWhat is Underfitting?\n\n### Response:\nUnderfitting: model too simple, performs poorly on both train and test"
    },
    {
        "instruction": "What is Too high?",
        "input": "",
        "output": "Too high: overshoots minimum",
        "text": "### Instruction:\nWhat is Too high?\n\n### Response:\nToo high: overshoots minimum"
    },
    {
        "instruction": "What is Duplicate ratio of a?",
        "input": "",
        "output": "Duplicate ratio of a: b is a^2:b^2",
        "text": "### Instruction:\nWhat is Duplicate ratio of a?\n\n### Response:\nDuplicate ratio of a: b is a^2:b^2"
    },
    {
        "instruction": "What is the significance/value of Article 280?",
        "input": "",
        "output": "Article 280: Finance Commission",
        "text": "### Instruction:\nWhat is the significance/value of Article 280?\n\n### Response:\nArticle 280: Finance Commission"
    },
    {
        "instruction": "What is LCM (Least Common Multiple)?",
        "input": "",
        "output": "LCM (Least Common Multiple): smallest number divisible by both numbers",
        "text": "### Instruction:\nWhat is LCM (Least Common Multiple)?\n\n### Response:\nLCM (Least Common Multiple): smallest number divisible by both numbers"
    },
    {
        "instruction": "What is Premium?",
        "input": "",
        "output": "Premium: 436 per year, Cover: 2 lakh",
        "text": "### Instruction:\nWhat is Premium?\n\n### Response:\nPremium: 436 per year, Cover: 2 lakh"
    },
    {
        "instruction": "What is Largest state by population?",
        "input": "",
        "output": "Largest state by population: Uttar Pradesh",
        "text": "### Instruction:\nWhat is Largest state by population?\n\n### Response:\nLargest state by population: Uttar Pradesh"
    },
    {
        "instruction": "What is If A is x% less than B, B is more than A by?",
        "input": "",
        "output": "If A is x% less than B, B is more than A by: (x / 100-x) x 100 percent",
        "text": "### Instruction:\nWhat is If A is x% less than B, B is more than A by?\n\n### Response:\nIf A is x% less than B, B is more than A by: (x / 100-x) x 100 percent"
    },
    {
        "instruction": "What is Two-factor authentication?",
        "input": "",
        "output": "Two-factor authentication: extra security layer beyond password",
        "text": "### Instruction:\nWhat is Two-factor authentication?\n\n### Response:\nTwo-factor authentication: extra security layer beyond password"
    },
    {
        "instruction": "What is the significance/value of Article 352?",
        "input": "",
        "output": "Article 352: National Emergency (armed rebellion or external aggression)",
        "text": "### Instruction:\nWhat is the significance/value of Article 352?\n\n### Response:\nArticle 352: National Emergency (armed rebellion or external aggression)"
    },
    {
        "instruction": "What is Deccan Plateau?",
        "input": "",
        "output": "Deccan Plateau: largest plateau in India",
        "text": "### Instruction:\nWhat is Deccan Plateau?\n\n### Response:\nDeccan Plateau: largest plateau in India"
    },
    {
        "instruction": "What is Article 14?",
        "input": "",
        "output": "Article 14: Equality before law",
        "text": "### Instruction:\nWhat is Article 14?\n\n### Response:\nArticle 14: Equality before law"
    },
    {
        "instruction": "What is Composite numbers?",
        "input": "",
        "output": "Composite numbers: more than 2 factors. Example: 4, 6, 8, 9",
        "text": "### Instruction:\nWhat is Composite numbers?\n\n### Response:\nComposite numbers: more than 2 factors. Example: 4, 6, 8, 9"
    },
    {
        "instruction": "What is What?",
        "input": "",
        "output": "What: AI-powered speech therapy platform for people with speech disorders",
        "text": "### Instruction:\nWhat is What?\n\n### Response:\nWhat: AI-powered speech therapy platform for people with speech disorders"
    },
    {
        "instruction": "What is Article 13?",
        "input": "",
        "output": "Article 13: Laws inconsistent with Fundamental Rights are void",
        "text": "### Instruction:\nWhat is Article 13?\n\n### Response:\nArticle 13: Laws inconsistent with Fundamental Rights are void"
    },
    {
        "instruction": "What is 29 July?",
        "input": "",
        "output": "29 July: National Tiger Day",
        "text": "### Instruction:\nWhat is 29 July?\n\n### Response:\n29 July: National Tiger Day"
    },
    {
        "instruction": "What is When compounded quarterly?",
        "input": "",
        "output": "When compounded quarterly: Rate = R/4, Time = 4T",
        "text": "### Instruction:\nWhat is When compounded quarterly?\n\n### Response:\nWhen compounded quarterly: Rate = R/4, Time = 4T"
    },
    {
        "instruction": "What is Doubtful Asset?",
        "input": "",
        "output": "Doubtful Asset: NPA for more than 12 months",
        "text": "### Instruction:\nWhat is Doubtful Asset?\n\n### Response:\nDoubtful Asset: NPA for more than 12 months"
    },
    {
        "instruction": "What is the formula for SECTION 7: AI/ML INTERVIEW PREPARATION?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for SECTION 7: AI/ML INTERVIEW PREPARATION?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is the formula for Profit = Selling Price - Cost Price?",
        "input": "",
        "output": "Loss% = (Loss / Cost Price) x 100",
        "text": "### Instruction:\nWhat is the formula for Profit = Selling Price - Cost Price?\n\n### Response:\nLoss% = (Loss / Cost Price) x 100"
    },
    {
        "instruction": "What is Database?",
        "input": "",
        "output": "Database: organized collection of data",
        "text": "### Instruction:\nWhat is Database?\n\n### Response:\nDatabase: organized collection of data"
    },
    {
        "instruction": "What is IPv4?",
        "input": "",
        "output": "IPv4: 32-bit address (4 groups of numbers, e.g., 192.168.1.1)",
        "text": "### Instruction:\nWhat is IPv4?\n\n### Response:\nIPv4: 32-bit address (4 groups of numbers, e.g., 192.168.1.1)"
    },
    {
        "instruction": "What is Circle?",
        "input": "",
        "output": "Circle: Area = pi x r^2, Circumference = 2 x pi x r",
        "text": "### Instruction:\nWhat is Circle?\n\n### Response:\nCircle: Area = pi x r^2, Circumference = 2 x pi x r"
    },
    {
        "instruction": "What is the formula for A > B means A is definitely greater than B?",
        "input": "",
        "output": "Complementary pairs: >, < and >=, <=",
        "text": "### Instruction:\nWhat is the formula for A > B means A is definitely greater than B?\n\n### Response:\nComplementary pairs: >, < and >=, <="
    },
    {
        "instruction": "What is HTTP?",
        "input": "",
        "output": "HTTP: HyperText Transfer Protocol - web browsing",
        "text": "### Instruction:\nWhat is HTTP?\n\n### Response:\nHTTP: HyperText Transfer Protocol - web browsing"
    },
    {
        "instruction": "What is the formula for Average Speed = Total Distance / Total Time?",
        "input": "",
        "output": "- Moving opposite direction: Time = (L1+L2) / (S1+S2)",
        "text": "### Instruction:\nWhat is the formula for Average Speed = Total Distance / Total Time?\n\n### Response:\n- Moving opposite direction: Time = (L1+L2) / (S1+S2)"
    },
    {
        "instruction": "What is the formula for Compound Interest = Amount - Principal?",
        "input": "",
        "output": "Difference between CI and SI for 2 years = P x (R/100)^2",
        "text": "### Instruction:\nWhat is the formula for Compound Interest = Amount - Principal?\n\n### Response:\nDifference between CI and SI for 2 years = P x (R/100)^2"
    },
    {
        "instruction": "What is IPv6?",
        "input": "",
        "output": "IPv6: 128-bit address (solves IPv4 exhaustion)",
        "text": "### Instruction:\nWhat is IPv6?\n\n### Response:\nIPv6: 128-bit address (solves IPv4 exhaustion)"
    },
    {
        "instruction": "What is the formula for Loss% = (Loss / Cost Price) x 100?",
        "input": "",
        "output": "If two items sold at same price, one at x% profit and one at x% loss, overall loss% = x^2/100",
        "text": "### Instruction:\nWhat is the formula for Loss% = (Loss / Cost Price) x 100?\n\n### Response:\nIf two items sold at same price, one at x% profit and one at x% loss, overall loss% = x^2/100"
    },
    {
        "instruction": "What is 26 January?",
        "input": "",
        "output": "26 January: Republic Day",
        "text": "### Instruction:\nWhat is 26 January?\n\n### Response:\n26 January: Republic Day"
    },
    {
        "instruction": "What is the significance/value of Article 21?",
        "input": "",
        "output": "Article 21: Protection of life and personal liberty",
        "text": "### Instruction:\nWhat is the significance/value of Article 21?\n\n### Response:\nArticle 21: Protection of life and personal liberty"
    },
    {
        "instruction": "What is Find opening sentence?",
        "input": "",
        "output": "Find opening sentence: introduces topic, no pronoun reference without antecedent",
        "text": "### Instruction:\nWhat is Find opening sentence?\n\n### Response:\nFind opening sentence: introduces topic, no pronoun reference without antecedent"
    },
    {
        "instruction": "What is All A are B?",
        "input": "",
        "output": "All A are B: every A is included in B",
        "text": "### Instruction:\nWhat is All A are B?\n\n### Response:\nAll A are B: every A is included in B"
    },
    {
        "instruction": "What is Letter coding?",
        "input": "",
        "output": "Letter coding: A=1, B=2, ..., Z=26 (or reverse Z=1)",
        "text": "### Instruction:\nWhat is Letter coding?\n\n### Response:\nLetter coding: A=1, B=2, ..., Z=26 (or reverse Z=1)"
    },
    {
        "instruction": "What is Longest river?",
        "input": "",
        "output": "Longest river: Ganga (but Indus is longest if including Pakistan part)",
        "text": "### Instruction:\nWhat is Longest river?\n\n### Response:\nLongest river: Ganga (but Indus is longest if including Pakistan part)"
    },
    {
        "instruction": "What are the key concepts in Reading Comprehension?",
        "input": "",
        "output": "Read passage carefully before questions\nIdentify main idea, supporting details, tone\nInference questions: answer not directly stated, must be logically drawn\nVocabulary in context: use surrounding words to find meaning\nTone can be: critical, appreciative, neutral, sarcastic, analytical",
        "text": "### Instruction:\nWhat are the key concepts in Reading Comprehension?\n\n### Response:\nRead passage carefully before questions\nIdentify main idea, supporting details, tone\nInference questions: answer not directly stated, must be logically drawn\nVocabulary in context: use surrounding words to find meaning\nTone can be: critical, appreciative, neutral, sarcastic, analytical"
    },
    {
        "instruction": "What is Early Stopping?",
        "input": "",
        "output": "Early Stopping: stop training when validation loss stops improving",
        "text": "### Instruction:\nWhat is Early Stopping?\n\n### Response:\nEarly Stopping: stop training when validation loss stops improving"
    },
    {
        "instruction": "What is ROM?",
        "input": "",
        "output": "ROM: Read Only Memory - permanent storage, non-volatile",
        "text": "### Instruction:\nWhat is ROM?\n\n### Response:\nROM: Read Only Memory - permanent storage, non-volatile"
    },
    {
        "instruction": "What is the formula for Probability always between 0 and 1?",
        "input": "",
        "output": "For mutually exclusive events: P(A or B) = P(A) + P(B)",
        "text": "### Instruction:\nWhat is the formula for Probability always between 0 and 1?\n\n### Response:\nFor mutually exclusive events: P(A or B) = P(A) + P(B)"
    },
    {
        "instruction": "What is the formula for handles long dependencies easily, much faster to train, but needs more data.?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for handles long dependencies easily, much faster to train, but needs more data.?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is Neural Network?",
        "input": "",
        "output": "Neural Network: layers of interconnected neurons",
        "text": "### Instruction:\nWhat is Neural Network?\n\n### Response:\nNeural Network: layers of interconnected neurons"
    },
    {
        "instruction": "What is GDP?",
        "input": "",
        "output": "GDP: Gross Domestic Product - total value of goods/services in a year",
        "text": "### Instruction:\nWhat is GDP?\n\n### Response:\nGDP: Gross Domestic Product - total value of goods/services in a year"
    },
    {
        "instruction": "What is SECTION 2?",
        "input": "",
        "output": "SECTION 2: REASONING ABILITY",
        "text": "### Instruction:\nWhat is SECTION 2?\n\n### Response:\nSECTION 2: REASONING ABILITY"
    },
    {
        "instruction": "Find the average of first 10 natural numbers.",
        "output": "Sum of first n natural numbers = n(n+1)/2 = 10x11/2 = 55. Average = 55/10 = 5.5",
        "text": "### Instruction:\nFind the average of first 10 natural numbers.\n\n### Response:\nSum of first n natural numbers = n(n+1)/2 = 10x11/2 = 55. Average = 55/10 = 5.5"
    },
    {
        "instruction": "What is 2. Unsupervised Learning?",
        "input": "",
        "output": "2. Unsupervised Learning: unlabeled data, find patterns (clustering, dimensionality reduction)",
        "text": "### Instruction:\nWhat is 2. Unsupervised Learning?\n\n### Response:\n2. Unsupervised Learning: unlabeled data, find patterns (clustering, dimensionality reduction)"
    },
    {
        "instruction": "What is Link sentences using?",
        "input": "",
        "output": "Link sentences using: pronouns, connectors, time sequence",
        "text": "### Instruction:\nWhat is Link sentences using?\n\n### Response:\nLink sentences using: pronouns, connectors, time sequence"
    },
    {
        "instruction": "What is Primary Key?",
        "input": "",
        "output": "Primary Key: unique identifier for each record",
        "text": "### Instruction:\nWhat is Primary Key?\n\n### Response:\nPrimary Key: unique identifier for each record"
    },
    {
        "instruction": "What is the formula for Steps for solving: draw family tree, use M/F to denote gender?",
        "input": "",
        "output": "Letter coding: A=1, B=2, ..., Z=26 (or reverse Z=1)",
        "text": "### Instruction:\nWhat is the formula for Steps for solving: draw family tree, use M/F to denote gender?\n\n### Response:\nLetter coding: A=1, B=2, ..., Z=26 (or reverse Z=1)"
    },
    {
        "instruction": "What is Fine-tuning?",
        "input": "",
        "output": "Fine-tuning: unfreeze and train all layers with small learning rate",
        "text": "### Instruction:\nWhat is Fine-tuning?\n\n### Response:\nFine-tuning: unfreeze and train all layers with small learning rate"
    },
    {
        "instruction": "What is Cache?",
        "input": "",
        "output": "Cache: fastest memory, stores frequently used data",
        "text": "### Instruction:\nWhat is Cache?\n\n### Response:\nCache: fastest memory, stores frequently used data"
    },
    {
        "instruction": "What is Malware?",
        "input": "",
        "output": "Malware: malicious software",
        "text": "### Instruction:\nWhat is Malware?\n\n### Response:\nMalware: malicious software"
    },
    {
        "instruction": "What is the formula for --- Compound Interest ---?",
        "input": "",
        "output": "Compound Interest = Amount - Principal",
        "text": "### Instruction:\nWhat is the formula for --- Compound Interest ---?\n\n### Response:\nCompound Interest = Amount - Principal"
    },
    {
        "instruction": "What is Active voice?",
        "input": "",
        "output": "Active voice: subject does action. Passive voice: subject receives action",
        "text": "### Instruction:\nWhat is Active voice?\n\n### Response:\nActive voice: subject does action. Passive voice: subject receives action"
    },
    {
        "instruction": "What are the key concepts in Government Schemes (Banking Related)?",
        "input": "",
        "output": "Pradhan Mantri Jan Dhan Yojana (PMJDY): financial inclusion scheme\n  - Zero balance account, RuPay debit card, overdraft facility\n  - Launched 28 August 2014\nPradhan Mantri Mudra Yojana: loans to small businesses\n  - Shishu: up to 50,000\n  - Kishore: 50,000 to 5 lakh\n  - Tarun: 5 lakh to 10 lakh\nPradhan Mantri Jeevan Jyoti Bima Yojana (PMJJBY): life insurance scheme\n  - Premium: 436 per year, Cover: 2 lakh\nPradhan Mantri Suraksha Bima Yojana (PMSBY): accident insurance\n  - Premium: 20 per year, Cover: 2 lakh\nAtal Pension Yojana (APY): pension scheme for unorganised sector\nStand-Up India: loans",
        "text": "### Instruction:\nWhat are the key concepts in Government Schemes (Banking Related)?\n\n### Response:\nPradhan Mantri Jan Dhan Yojana (PMJDY): financial inclusion scheme\n  - Zero balance account, RuPay debit card, overdraft facility\n  - Launched 28 August 2014\nPradhan Mantri Mudra Yojana: loans to small businesses\n  - Shishu: up to 50,000\n  - Kishore: 50,000 to 5 lakh\n  - Tarun: 5 lakh to 10 lakh\nPradhan Mantri Jeevan Jyoti Bima Yojana (PMJJBY): life insurance scheme\n  - Premium: 436 per year, Cover: 2 lakh\nPradhan Mantri Suraksha Bima Yojana (PMSBY): accident insurance\n  - Premium: 20 per year, Cover: 2 lakh\nAtal Pension Yojana (APY): pension scheme for unorganised sector\nStand-Up India: loans"
    },
    {
        "instruction": "What is the formula for Profit% = (Profit / Cost Price) x 100?",
        "input": "",
        "output": "Cost Price = Selling Price / (1 + Profit%/100)",
        "text": "### Instruction:\nWhat is the formula for Profit% = (Profit / Cost Price) x 100?\n\n### Response:\nCost Price = Selling Price / (1 + Profit%/100)"
    },
    {
        "instruction": "What is R-CNN family?",
        "input": "",
        "output": "R-CNN family: region-based detection",
        "text": "### Instruction:\nWhat is R-CNN family?\n\n### Response:\nR-CNN family: region-based detection"
    },
    {
        "instruction": "What is HTTPS?",
        "input": "",
        "output": "HTTPS: Secure version of HTTP",
        "text": "### Instruction:\nWhat is HTTPS?\n\n### Response:\nHTTPS: Secure version of HTTP"
    },
    {
        "instruction": "What is the formula for --- Profit and Loss ---?",
        "input": "",
        "output": "Loss = Cost Price - Selling Price",
        "text": "### Instruction:\nWhat is the formula for --- Profit and Loss ---?\n\n### Response:\nLoss = Cost Price - Selling Price"
    },
    {
        "instruction": "How do you handle imbalanced datasets?",
        "input": "",
        "output": "Oversampling minority class (SMOTE), undersampling majority class,\nclass weights in loss function, use F1/AUC instead of accuracy.",
        "text": "### Instruction:\nHow do you handle imbalanced datasets?\n\n### Response:\nOversampling minority class (SMOTE), undersampling majority class,\nclass weights in loss function, use F1/AUC instead of accuracy."
    },
    {
        "instruction": "What are the key concepts in Simple Interest?",
        "input": "",
        "output": "Simple Interest (SI) = (Principal x Rate x Time) / 100\nAmount = Principal + SI\nPrincipal = (SI x 100) / (Rate x Time)\nRate = (SI x 100) / (Principal x Time)\nTime = (SI x 100) / (Principal x Rate)",
        "text": "### Instruction:\nWhat are the key concepts in Simple Interest?\n\n### Response:\nSimple Interest (SI) = (Principal x Rate x Time) / 100\nAmount = Principal + SI\nPrincipal = (SI x 100) / (Rate x Time)\nRate = (SI x 100) / (Principal x Time)\nTime = (SI x 100) / (Principal x Rate)"
    },
    {
        "instruction": "What is KYC?",
        "input": "",
        "output": "KYC: Know Your Customer - identity verification process",
        "text": "### Instruction:\nWhat is KYC?\n\n### Response:\nKYC: Know Your Customer - identity verification process"
    },
    {
        "instruction": "What is Fix?",
        "input": "",
        "output": "Fix: remove duplicates, keep single version of each",
        "text": "### Instruction:\nWhat is Fix?\n\n### Response:\nFix: remove duplicates, keep single version of each"
    },
    {
        "instruction": "What is 3. Reinforcement Learning?",
        "input": "",
        "output": "3. Reinforcement Learning: agent learns by reward/punishment",
        "text": "### Instruction:\nWhat is 3. Reinforcement Learning?\n\n### Response:\n3. Reinforcement Learning: agent learns by reward/punishment"
    },
    {
        "instruction": "What is classifier. Fine-tuning?",
        "input": "",
        "output": "classifier. Fine-tuning: unfreeze and train all layers with small learning rate.",
        "text": "### Instruction:\nWhat is classifier. Fine-tuning?\n\n### Response:\nclassifier. Fine-tuning: unfreeze and train all layers with small learning rate."
    },
    {
        "instruction": "What is Highest peak?",
        "input": "",
        "output": "Highest peak: Kangchenjunga (Sikkim) - 3rd highest in world",
        "text": "### Instruction:\nWhat is Highest peak?\n\n### Response:\nHighest peak: Kangchenjunga (Sikkim) - 3rd highest in world"
    },
    {
        "instruction": "What is the formula for If A is x% less than B, B is more than A by: (x / 100-x) x 100 percent?",
        "input": "",
        "output": "Profit = Selling Price - Cost Price",
        "text": "### Instruction:\nWhat is the formula for If A is x% less than B, B is more than A by: (x / 100-x) x 100 percent?\n\n### Response:\nProfit = Selling Price - Cost Price"
    },
    {
        "instruction": "What is the formula for Circle: Area = pi x r^2, Circumference = 2 x pi x r?",
        "input": "",
        "output": "Sphere: Volume = 4/3 x pi x r^3, Surface Area = 4 x pi x r^2",
        "text": "### Instruction:\nWhat is the formula for Circle: Area = pi x r^2, Circumference = 2 x pi x r?\n\n### Response:\nSphere: Volume = 4/3 x pi x r^3, Surface Area = 4 x pi x r^2"
    },
    {
        "instruction": "What is Kernel trick?",
        "input": "",
        "output": "Kernel trick: maps data to higher dimension",
        "text": "### Instruction:\nWhat is Kernel trick?\n\n### Response:\nKernel trick: maps data to higher dimension"
    },
    {
        "instruction": "What are the key concepts in Compound Interest?",
        "input": "",
        "output": "Amount = Principal x (1 + Rate/100)^Time\nCompound Interest = Amount - Principal\nWhen compounded half-yearly: Rate = R/2, Time = 2T\nWhen compounded quarterly: Rate = R/4, Time = 4T\nDifference between CI and SI for 2 years = P x (R/100)^2\nDifference between CI and SI for 3 years = P x (R/100)^2 x (R/100 + 3)",
        "text": "### Instruction:\nWhat are the key concepts in Compound Interest?\n\n### Response:\nAmount = Principal x (1 + Rate/100)^Time\nCompound Interest = Amount - Principal\nWhen compounded half-yearly: Rate = R/2, Time = 2T\nWhen compounded quarterly: Rate = R/4, Time = 4T\nDifference between CI and SI for 2 years = P x (R/100)^2\nDifference between CI and SI for 3 years = P x (R/100)^2 x (R/100 + 3)"
    },
    {
        "instruction": "What is regularization and why is it used?",
        "input": "",
        "output": "Regularization adds penalty to loss function to prevent overfitting.\nL1 (Lasso) produces sparse models. L2 (Ridge) prevents large weights.\nDropout randomly deactivates neurons. All reduce model complexity.",
        "text": "### Instruction:\nWhat is regularization and why is it used?\n\n### Response:\nRegularization adds penalty to loss function to prevent overfitting.\nL1 (Lasso) produces sparse models. L2 (Ridge) prevents large weights.\nDropout randomly deactivates neurons. All reduce model complexity."
    },
    {
        "instruction": "What is Cube?",
        "input": "",
        "output": "Cube: Volume = a^3, Surface Area = 6a^2",
        "text": "### Instruction:\nWhat is Cube?\n\n### Response:\nCube: Volume = a^3, Surface Area = 6a^2"
    },
    {
        "instruction": "Query, K: Key, V: Value\n  - Multi-head attention: run attention multiple times in parallel\n  - Positional encoding: adds position information\n  - Feed-forward network: applied to each position\n  - Layer normalization: stabilizes training\n\nBERT: Bidirectional Encoder, pre-trained on masked language modeling\nGPT: Generative Pre-trained Transformer, autoregressive generation\nMistral: efficient transformer, sliding window attention, grouped query attention\n\n--- NLP Concepts ---\nTokenization: splitting text into tokens (words, subwords, characters)\nEmbedding: mapping tokens to dense vectors (Word2Vec, GloVe, FastText)\nTF-IDF: Term Frequency - Inverse Document Frequency\n  - TF: how often word appears in document\n  - IDF: how rare word is across all documents\nSentiment Analysis: classify text as positive/negative/neutral\nNamed Entity Recognition (NER): identify names, places, organizations\nText Classification: assign categories to text\n\nUsed in Parth's Projects:\n  - NeuroSpeech: OpenAI Whisper for speech-to-text (transformer-based)\n  - IPA phonetic analysis for pronunciation feedback\n\n--- Computer Vision ---\nImage preprocessing: resize, normalize, augment\nData augmentation: flip, rotate, zoom, shift, brightness change\n  - Used in Drive-Awake training\nTransfer Learning: use pre-trained model weights on new task\n  - Feature extraction: freeze pretrained layers, train only final layers\n  - Fine-tuning: unfreeze and train all layers with small learning rate\nObject Detection: locate and classify objects in image\n  - YOLO: You Only Look Once, real-time detection\n  - R-CNN family: region-based detection\n\nOpenCV:\n  - cv2.imread(): read image\n  - cv2.cvtColor(): convert color space (BGR to RGB, BGR to Gray)\n  - cv2.resize(): resize image\n  - cv2.CascadeClassifier(): Haar cascade for face/eye detection\n  - Eye Aspect Ratio (EAR) = (||p2-p6|| + ||p3-p5||) / (2 * ||p1-p4||)\n  - EAR < 0.25 for consecutive frames = drowsiness detected\n\n--- Evaluation Metrics ---\nClassification:\n  - Accuracy = (TP + TN) / (TP + TN + FP + FN)\n  - Precision = TP / (TP + FP) - out of predicted positives, how many are correct\n  - Recall = TP / (TP + FN) - out of actual positives, how many found\n  - F1 Score = 2 x (Precision x Recall) / (Precision + Recall)\n  - ROC-AUC: area under ROC curve, higher is better\n  - Confusion Matrix: TP, TN, FP, FN\n\nRegression:\n  - MAE (Mean Absolute Error): average absolute difference\n  - MSE (Mean Squared Error): average squared difference\n  - RMSE: square root of MSE\n  - R-squared: proportion of variance explained by model (0 to 1)\n\nClustering:\n  - Silhouette Score: how similar object is to its cluster vs other clusters\n  - Inertia (within-cluster sum of squares): lower is better\n\n--- Python for ML ---\nNumPy: numerical computing, arrays, matrix operations\n  - np.array(), np.reshape(), np.dot(), np.mean(), np.std()\n  - Broadcasting: operations on different shaped arrays\nPandas: data manipulation and analysis\n  - pd.read_csv(), df.head(), df.describe(), df.dropna(), df.fillna()\n  - df.groupby(), df.merge(), df.pivot_table()\nMatplotlib/Seaborn: data visualization\nScikit-learn: ML algorithms, preprocessing, evaluation\n  - train_test_split, StandardScaler, LabelEncoder\n  - KMeans, RandomForestClassifier, LinearRegression\nTensorFlow/Keras: deep learning framework\n  - model.fit(), model.predict(), model.evaluate()\n  - Sequential, Dense, Conv2D, MaxPooling2D, Dropout layers\nPyTorch: deep learning framework, dynamic computation graphs\n  - torch.tensor(), nn.Module, forward(), backward()\n  - DataLoader, Dataset\n\n--- MLOps and Deployment ---\nModel Serialization: saving trained models\n  - TensorFlow: model.save(), tf.keras.models.load_model()\n  - PyTorch: torch.save(), torch.load()\n  - Scikit-learn: pickle, joblib\nFastAPI: fast web framework for building ML APIs (Parth uses this)\n  - @app.post(), @app.get(), Pydantic models for validation\n  - Used in Drive-Awake for drowsiness detection API\nDocker: containerize ML applications\n  - Dockerfile, docker build, docker run\n  - FROM, RUN, COPY, CMD instructions\nHuggingFace: platform for sharing ML models and spaces\n  - Parth's NeuroSpeech is deployed on HuggingFace Spaces\n  - transformers library, pipeline()\nSupabase: open source Firebase alternative\n  - Authentication, PostgreSQL database, Edge Functions, Realtime\n  - Used extensively in Parth's NeuroSpeech project\n\n--- Common Interview Questions ---\nQ: What is the difference between supervised and unsupervised learning?",
        "input": "",
        "output": "Supervised learning uses labeled data to predict outputs (classification/regression).\nUnsupervised learning finds hidden patterns in unlabeled data (clustering/dimensionality reduction).",
        "text": "### Instruction:\nQuery, K: Key, V: Value\n  - Multi-head attention: run attention multiple times in parallel\n  - Positional encoding: adds position information\n  - Feed-forward network: applied to each position\n  - Layer normalization: stabilizes training\n\nBERT: Bidirectional Encoder, pre-trained on masked language modeling\nGPT: Generative Pre-trained Transformer, autoregressive generation\nMistral: efficient transformer, sliding window attention, grouped query attention\n\n--- NLP Concepts ---\nTokenization: splitting text into tokens (words, subwords, characters)\nEmbedding: mapping tokens to dense vectors (Word2Vec, GloVe, FastText)\nTF-IDF: Term Frequency - Inverse Document Frequency\n  - TF: how often word appears in document\n  - IDF: how rare word is across all documents\nSentiment Analysis: classify text as positive/negative/neutral\nNamed Entity Recognition (NER): identify names, places, organizations\nText Classification: assign categories to text\n\nUsed in Parth's Projects:\n  - NeuroSpeech: OpenAI Whisper for speech-to-text (transformer-based)\n  - IPA phonetic analysis for pronunciation feedback\n\n--- Computer Vision ---\nImage preprocessing: resize, normalize, augment\nData augmentation: flip, rotate, zoom, shift, brightness change\n  - Used in Drive-Awake training\nTransfer Learning: use pre-trained model weights on new task\n  - Feature extraction: freeze pretrained layers, train only final layers\n  - Fine-tuning: unfreeze and train all layers with small learning rate\nObject Detection: locate and classify objects in image\n  - YOLO: You Only Look Once, real-time detection\n  - R-CNN family: region-based detection\n\nOpenCV:\n  - cv2.imread(): read image\n  - cv2.cvtColor(): convert color space (BGR to RGB, BGR to Gray)\n  - cv2.resize(): resize image\n  - cv2.CascadeClassifier(): Haar cascade for face/eye detection\n  - Eye Aspect Ratio (EAR) = (||p2-p6|| + ||p3-p5||) / (2 * ||p1-p4||)\n  - EAR < 0.25 for consecutive frames = drowsiness detected\n\n--- Evaluation Metrics ---\nClassification:\n  - Accuracy = (TP + TN) / (TP + TN + FP + FN)\n  - Precision = TP / (TP + FP) - out of predicted positives, how many are correct\n  - Recall = TP / (TP + FN) - out of actual positives, how many found\n  - F1 Score = 2 x (Precision x Recall) / (Precision + Recall)\n  - ROC-AUC: area under ROC curve, higher is better\n  - Confusion Matrix: TP, TN, FP, FN\n\nRegression:\n  - MAE (Mean Absolute Error): average absolute difference\n  - MSE (Mean Squared Error): average squared difference\n  - RMSE: square root of MSE\n  - R-squared: proportion of variance explained by model (0 to 1)\n\nClustering:\n  - Silhouette Score: how similar object is to its cluster vs other clusters\n  - Inertia (within-cluster sum of squares): lower is better\n\n--- Python for ML ---\nNumPy: numerical computing, arrays, matrix operations\n  - np.array(), np.reshape(), np.dot(), np.mean(), np.std()\n  - Broadcasting: operations on different shaped arrays\nPandas: data manipulation and analysis\n  - pd.read_csv(), df.head(), df.describe(), df.dropna(), df.fillna()\n  - df.groupby(), df.merge(), df.pivot_table()\nMatplotlib/Seaborn: data visualization\nScikit-learn: ML algorithms, preprocessing, evaluation\n  - train_test_split, StandardScaler, LabelEncoder\n  - KMeans, RandomForestClassifier, LinearRegression\nTensorFlow/Keras: deep learning framework\n  - model.fit(), model.predict(), model.evaluate()\n  - Sequential, Dense, Conv2D, MaxPooling2D, Dropout layers\nPyTorch: deep learning framework, dynamic computation graphs\n  - torch.tensor(), nn.Module, forward(), backward()\n  - DataLoader, Dataset\n\n--- MLOps and Deployment ---\nModel Serialization: saving trained models\n  - TensorFlow: model.save(), tf.keras.models.load_model()\n  - PyTorch: torch.save(), torch.load()\n  - Scikit-learn: pickle, joblib\nFastAPI: fast web framework for building ML APIs (Parth uses this)\n  - @app.post(), @app.get(), Pydantic models for validation\n  - Used in Drive-Awake for drowsiness detection API\nDocker: containerize ML applications\n  - Dockerfile, docker build, docker run\n  - FROM, RUN, COPY, CMD instructions\nHuggingFace: platform for sharing ML models and spaces\n  - Parth's NeuroSpeech is deployed on HuggingFace Spaces\n  - transformers library, pipeline()\nSupabase: open source Firebase alternative\n  - Authentication, PostgreSQL database, Edge Functions, Realtime\n  - Used extensively in Parth's NeuroSpeech project\n\n--- Common Interview Questions ---\nQ: What is the difference between supervised and unsupervised learning?\n\n### Response:\nSupervised learning uses labeled data to predict outputs (classification/regression).\nUnsupervised learning finds hidden patterns in unlabeled data (clustering/dimensionality reduction)."
    },
    {
        "instruction": "What is the formula for Time = (SI x 100) / (Principal x Rate)?",
        "input": "",
        "output": "Amount = Principal x (1 + Rate/100)^Time",
        "text": "### Instruction:\nWhat is the formula for Time = (SI x 100) / (Principal x Rate)?\n\n### Response:\nAmount = Principal x (1 + Rate/100)^Time"
    },
    {
        "instruction": "What is the formula for Simple Interest (SI) = (Principal x Rate x Time) / 100?",
        "input": "",
        "output": "Rate = (SI x 100) / (Principal x Time)",
        "text": "### Instruction:\nWhat is the formula for Simple Interest (SI) = (Principal x Rate x Time) / 100?\n\n### Response:\nRate = (SI x 100) / (Principal x Time)"
    },
    {
        "instruction": "What is Mixed series?",
        "input": "",
        "output": "Mixed series: combine two or more patterns",
        "text": "### Instruction:\nWhat is Mixed series?\n\n### Response:\nMixed series: combine two or more patterns"
    },
    {
        "instruction": "What is the significance/value of Kishore?",
        "input": "",
        "output": "Kishore: 50,000 to 5 lakh",
        "text": "### Instruction:\nWhat is the significance/value of Kishore?\n\n### Response:\nKishore: 50,000 to 5 lakh"
    },
    {
        "instruction": "What is Adam Optimizer?",
        "input": "",
        "output": "Adam Optimizer: adaptive learning rates, most popular",
        "text": "### Instruction:\nWhat is Adam Optimizer?\n\n### Response:\nAdam Optimizer: adaptive learning rates, most popular"
    },
    {
        "instruction": "What is the formula for - Moving opposite direction: Time = (L1+L2) / (S1+S2)?",
        "input": "",
        "output": "Relative speed (same direction) = S1 - S2",
        "text": "### Instruction:\nWhat is the formula for - Moving opposite direction: Time = (L1+L2) / (S1+S2)?\n\n### Response:\nRelative speed (same direction) = S1 - S2"
    },
    {
        "instruction": "What are the key concepts in Types of Banks?",
        "input": "",
        "output": "Central Bank: RBI (regulates all banks)\nCommercial Banks: SBI, PNB, HDFC, ICICI, Axis\n  - Public sector: majority government owned (SBI, BOB, PNB)\n  - Private sector: majority private owned (HDFC, ICICI, Axis)\n  - Foreign banks: headquartered abroad (Citibank, HSBC)\nCooperative Banks: serve specific communities\nRegional Rural Banks (RRBs): serve rural areas\nSmall Finance Banks: serve small borrowers and unserved sections\nPayment Banks: can accept deposits up to 2 lakh, cannot lend\nDevelopment Banks: NABARD, NHB, SIDBI, EXIM Bank",
        "text": "### Instruction:\nWhat are the key concepts in Types of Banks?\n\n### Response:\nCentral Bank: RBI (regulates all banks)\nCommercial Banks: SBI, PNB, HDFC, ICICI, Axis\n  - Public sector: majority government owned (SBI, BOB, PNB)\n  - Private sector: majority private owned (HDFC, ICICI, Axis)\n  - Foreign banks: headquartered abroad (Citibank, HSBC)\nCooperative Banks: serve specific communities\nRegional Rural Banks (RRBs): serve rural areas\nSmall Finance Banks: serve small borrowers and unserved sections\nPayment Banks: can accept deposits up to 2 lakh, cannot lend\nDevelopment Banks: NABARD, NHB, SIDBI, EXIM Bank"
    },
    {
        "instruction": "What is Repo Rate?",
        "input": "",
        "output": "Repo Rate: rate at which RBI lends to commercial banks",
        "text": "### Instruction:\nWhat is Repo Rate?\n\n### Response:\nRepo Rate: rate at which RBI lends to commercial banks"
    },
    {
        "instruction": "What is SECTION 4?",
        "input": "",
        "output": "SECTION 4: BANKING AND FINANCIAL AWARENESS",
        "text": "### Instruction:\nWhat is SECTION 4?\n\n### Response:\nSECTION 4: BANKING AND FINANCIAL AWARENESS"
    },
    {
        "instruction": "What are the key concepts in Evaluation Metrics?",
        "input": "",
        "output": "Classification:\n  - Accuracy = (TP + TN) / (TP + TN + FP + FN)\n  - Precision = TP / (TP + FP) - out of predicted positives, how many are correct\n  - Recall = TP / (TP + FN) - out of actual positives, how many found\n  - F1 Score = 2 x (Precision x Recall) / (Precision + Recall)\n  - ROC-AUC: area under ROC curve, higher is better\n  - Confusion Matrix: TP, TN, FP, FN\n\nRegression:\n  - MAE (Mean Absolute Error): average absolute difference\n  - MSE (Mean Squared Error): average squared difference\n  - RMSE: square root of MSE\n  - R-squared: proportion of variance explained by model (0 to 1)\n\nClusteri",
        "text": "### Instruction:\nWhat are the key concepts in Evaluation Metrics?\n\n### Response:\nClassification:\n  - Accuracy = (TP + TN) / (TP + TN + FP + FN)\n  - Precision = TP / (TP + FP) - out of predicted positives, how many are correct\n  - Recall = TP / (TP + FN) - out of actual positives, how many found\n  - F1 Score = 2 x (Precision x Recall) / (Precision + Recall)\n  - ROC-AUC: area under ROC curve, higher is better\n  - Confusion Matrix: TP, TN, FP, FN\n\nRegression:\n  - MAE (Mean Absolute Error): average absolute difference\n  - MSE (Mean Squared Error): average squared difference\n  - RMSE: square root of MSE\n  - R-squared: proportion of variance explained by model (0 to 1)\n\nClusteri"
    },
    {
        "instruction": "What is SECTION 6?",
        "input": "",
        "output": "SECTION 6: COMPUTER AWARENESS (FOR BANKING EXAMS)",
        "text": "### Instruction:\nWhat is SECTION 6?\n\n### Response:\nSECTION 6: COMPUTER AWARENESS (FOR BANKING EXAMS)"
    },
    {
        "instruction": "What is Article 17?",
        "input": "",
        "output": "Article 17: Abolition of untouchability",
        "text": "### Instruction:\nWhat is Article 17?\n\n### Response:\nArticle 17: Abolition of untouchability"
    },
    {
        "instruction": "What is Five Year Plans?",
        "input": "",
        "output": "Five Year Plans: replaced by 3-year action plan + 7-year strategy",
        "text": "### Instruction:\nWhat is Five Year Plans?\n\n### Response:\nFive Year Plans: replaced by 3-year action plan + 7-year strategy"
    },
    {
        "instruction": "What is the significance/value of Article 20?",
        "input": "",
        "output": "Article 20: Protection against self-incrimination",
        "text": "### Instruction:\nWhat is the significance/value of Article 20?\n\n### Response:\nArticle 20: Protection against self-incrimination"
    },
    {
        "instruction": "What is Silhouette Score?",
        "input": "",
        "output": "Silhouette Score: how similar object is to its cluster vs other clusters",
        "text": "### Instruction:\nWhat is Silhouette Score?\n\n### Response:\nSilhouette Score: how similar object is to its cluster vs other clusters"
    },
    {
        "instruction": "What is the formula for Square: Area = side^2, Perimeter = 4 x side?",
        "input": "",
        "output": "Circle: Area = pi x r^2, Circumference = 2 x pi x r",
        "text": "### Instruction:\nWhat is the formula for Square: Area = side^2, Perimeter = 4 x side?\n\n### Response:\nCircle: Area = pi x r^2, Circumference = 2 x pi x r"
    },
    {
        "instruction": "What are the key concepts in Direction and Distance?",
        "input": "",
        "output": "Starting direction: North\nTurning right from North = East\nTurning left from North = West\nAfter 180 degree turn from North = South\nShadow at sunrise: on West side\nShadow at sunset: on East side\nShortest distance: Pythagoras theorem for perpendicular paths",
        "text": "### Instruction:\nWhat are the key concepts in Direction and Distance?\n\n### Response:\nStarting direction: North\nTurning right from North = East\nTurning left from North = West\nAfter 180 degree turn from North = South\nShadow at sunrise: on West side\nShadow at sunset: on East side\nShortest distance: Pythagoras theorem for perpendicular paths"
    },
    {
        "instruction": "What is the significance/value of AlexNet?",
        "input": "",
        "output": "AlexNet: won ImageNet 2012, made deep learning popular",
        "text": "### Instruction:\nWhat is the significance/value of AlexNet?\n\n### Response:\nAlexNet: won ImageNet 2012, made deep learning popular"
    },
    {
        "instruction": "What is TF-IDF?",
        "input": "",
        "output": "TF-IDF: Term Frequency - Inverse Document Frequency",
        "text": "### Instruction:\nWhat is TF-IDF?\n\n### Response:\nTF-IDF: Term Frequency - Inverse Document Frequency"
    },
    {
        "instruction": "What is Revenue deficit?",
        "input": "",
        "output": "Revenue deficit: revenue expenditure exceeds revenue receipts",
        "text": "### Instruction:\nWhat is Revenue deficit?\n\n### Response:\nRevenue deficit: revenue expenditure exceeds revenue receipts"
    },
    {
        "instruction": "What is Subject-verb agreement?",
        "input": "",
        "output": "Subject-verb agreement: singular subject takes singular verb",
        "text": "### Instruction:\nWhat is Subject-verb agreement?\n\n### Response:\nSubject-verb agreement: singular subject takes singular verb"
    },
    {
        "instruction": "What is Live?",
        "input": "",
        "output": "Live: https://huggingface.co/spaces/parthbijpuriya/neurospeech",
        "text": "### Instruction:\nWhat is Live?\n\n### Response:\nLive: https://huggingface.co/spaces/parthbijpuriya/neurospeech"
    },
    {
        "instruction": "What is Gini Impurity?",
        "input": "",
        "output": "Gini Impurity: measures split quality",
        "text": "### Instruction:\nWhat is Gini Impurity?\n\n### Response:\nGini Impurity: measures split quality"
    },
    {
        "instruction": "What is Geometric series?",
        "input": "",
        "output": "Geometric series: ratio is constant (2,4,8,16...)",
        "text": "### Instruction:\nWhat is Geometric series?\n\n### Response:\nGeometric series: ratio is constant (2,4,8,16...)"
    },
    {
        "instruction": "What is Shadow at sunrise?",
        "input": "",
        "output": "Shadow at sunrise: on West side",
        "text": "### Instruction:\nWhat is Shadow at sunrise?\n\n### Response:\nShadow at sunrise: on West side"
    },
    {
        "instruction": "What is the formula for SECTION 1: QUANTITATIVE APTITUDE?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for SECTION 1: QUANTITATIVE APTITUDE?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is India's GDP growth rate?",
        "input": "",
        "output": "India's GDP growth rate: around 6-7% annually",
        "text": "### Instruction:\nWhat is India's GDP growth rate?\n\n### Response:\nIndia's GDP growth rate: around 6-7% annually"
    },
    {
        "instruction": "What is SEBI?",
        "input": "",
        "output": "SEBI: Securities and Exchange Board of India - regulates stock markets",
        "text": "### Instruction:\nWhat is SEBI?\n\n### Response:\nSEBI: Securities and Exchange Board of India - regulates stock markets"
    },
    {
        "instruction": "What are the key concepts in Probability?",
        "input": "",
        "output": "Probability = Favourable outcomes / Total outcomes\nProbability always between 0 and 1\nP(A) + P(not A) = 1\nP(A or B) = P(A) + P(B) - P(A and B)\nFor mutually exclusive events: P(A or B) = P(A) + P(B)\nFor independent events: P(A and B) = P(A) x P(B)\nCombination nCr = n! / (r! x (n-r)!)\nPermutation nPr = n! / (n-r)!\n\n================================================================================\nSECTION 2: REASONING ABILITY\n================================================================================",
        "text": "### Instruction:\nWhat are the key concepts in Probability?\n\n### Response:\nProbability = Favourable outcomes / Total outcomes\nProbability always between 0 and 1\nP(A) + P(not A) = 1\nP(A or B) = P(A) + P(B) - P(A and B)\nFor mutually exclusive events: P(A or B) = P(A) + P(B)\nFor independent events: P(A and B) = P(A) x P(B)\nCombination nCr = n! / (r! x (n-r)!)\nPermutation nPr = n! / (n-r)!\n\n================================================================================\nSECTION 2: REASONING ABILITY\n================================================================================"
    },
    {
        "instruction": "What is Article 1?",
        "input": "",
        "output": "Article 1: India shall be a Union of States",
        "text": "### Instruction:\nWhat is Article 1?\n\n### Response:\nArticle 1: India shall be a Union of States"
    },
    {
        "instruction": "What is Naive Bayes?",
        "input": "",
        "output": "Naive Bayes: probabilistic classifier, assumes feature independence",
        "text": "### Instruction:\nWhat is Naive Bayes?\n\n### Response:\nNaive Bayes: probabilistic classifier, assumes feature independence"
    },
    {
        "instruction": "What is India?",
        "input": "",
        "output": "India: 7th largest country by area, 2nd most populous",
        "text": "### Instruction:\nWhat is India?\n\n### Response:\nIndia: 7th largest country by area, 2nd most populous"
    },
    {
        "instruction": "What is SECTION 8?",
        "input": "",
        "output": "SECTION 8: PARTH'S PROJECT SUMMARIES (FOR INTERVIEWS)",
        "text": "### Instruction:\nWhat is SECTION 8?\n\n### Response:\nSECTION 8: PARTH'S PROJECT SUMMARIES (FOR INTERVIEWS)"
    },
    {
        "instruction": "What is Square?",
        "input": "",
        "output": "Square: Area = side^2, Perimeter = 4 x side",
        "text": "### Instruction:\nWhat is Square?\n\n### Response:\nSquare: Area = side^2, Perimeter = 4 x side"
    },
    {
        "instruction": "What is Regional Rural Banks (RRBs)?",
        "input": "",
        "output": "Regional Rural Banks (RRBs): serve rural areas",
        "text": "### Instruction:\nWhat is Regional Rural Banks (RRBs)?\n\n### Response:\nRegional Rural Banks (RRBs): serve rural areas"
    },
    {
        "instruction": "What is the significance/value of Payment Banks?",
        "input": "",
        "output": "Payment Banks: can accept deposits up to 2 lakh, cannot lend",
        "text": "### Instruction:\nWhat is the significance/value of Payment Banks?\n\n### Response:\nPayment Banks: can accept deposits up to 2 lakh, cannot lend"
    },
    {
        "instruction": "What are the key concepts in Blood Relations?",
        "input": "",
        "output": "Father's father = Paternal grandfather\nMother's father = Maternal grandfather\nFather's brother = Uncle, Father's sister = Aunt\nHusband's mother = Mother-in-law\nSon's wife = Daughter-in-law\nBrother's son = Nephew, Brother's daughter = Niece\nSteps for solving: draw family tree, use M/F to denote gender",
        "text": "### Instruction:\nWhat are the key concepts in Blood Relations?\n\n### Response:\nFather's father = Paternal grandfather\nMother's father = Maternal grandfather\nFather's brother = Uncle, Father's sister = Aunt\nHusband's mother = Mother-in-law\nSon's wife = Daughter-in-law\nBrother's son = Nephew, Brother's daughter = Niece\nSteps for solving: draw family tree, use M/F to denote gender"
    },
    {
        "instruction": "What is Too low?",
        "input": "",
        "output": "Too low: very slow convergence",
        "text": "### Instruction:\nWhat is Too low?\n\n### Response:\nToo low: very slow convergence"
    },
    {
        "instruction": "What is Payment Banks?",
        "input": "",
        "output": "Payment Banks: can accept deposits up to 2 lakh, cannot lend",
        "text": "### Instruction:\nWhat is Payment Banks?\n\n### Response:\nPayment Banks: can accept deposits up to 2 lakh, cannot lend"
    },
    {
        "instruction": "What is Named Entity Recognition (NER)?",
        "input": "",
        "output": "Named Entity Recognition (NER): identify names, places, organizations",
        "text": "### Instruction:\nWhat is Named Entity Recognition (NER)?\n\n### Response:\nNamed Entity Recognition (NER): identify names, places, organizations"
    },
    {
        "instruction": "What is Current Account?",
        "input": "",
        "output": "Current Account: for businesses, no interest, unlimited transactions",
        "text": "### Instruction:\nWhat is Current Account?\n\n### Response:\nCurrent Account: for businesses, no interest, unlimited transactions"
    },
    {
        "instruction": "What is 18 April?",
        "input": "",
        "output": "18 April: World Heritage Day",
        "text": "### Instruction:\nWhat is 18 April?\n\n### Response:\n18 April: World Heritage Day"
    },
    {
        "instruction": "What is Kernels?",
        "input": "",
        "output": "Kernels: Linear, RBF (Radial Basis Function), Polynomial",
        "text": "### Instruction:\nWhat is Kernels?\n\n### Response:\nKernels: Linear, RBF (Radial Basis Function), Polynomial"
    },
    {
        "instruction": "What is Variance?",
        "input": "",
        "output": "Variance: error from sensitivity to small fluctuations in training data",
        "text": "### Instruction:\nWhat is Variance?\n\n### Response:\nVariance: error from sensitivity to small fluctuations in training data"
    },
    {
        "instruction": "What is the formula for Amount = Principal + SI?",
        "input": "",
        "output": "Time = (SI x 100) / (Principal x Rate)",
        "text": "### Instruction:\nWhat is the formula for Amount = Principal + SI?\n\n### Response:\nTime = (SI x 100) / (Principal x Rate)"
    },
    {
        "instruction": "What is Stand-Up India?",
        "input": "",
        "output": "Stand-Up India: loans to SC/ST and women entrepreneurs",
        "text": "### Instruction:\nWhat is Stand-Up India?\n\n### Response:\nStand-Up India: loans to SC/ST and women entrepreneurs"
    },
    {
        "instruction": "What is the formula for SECTION 3: ENGLISH LANGUAGE?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for SECTION 3: ENGLISH LANGUAGE?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is Substandard Asset?",
        "input": "",
        "output": "Substandard Asset: NPA for less than 12 months",
        "text": "### Instruction:\nWhat is Substandard Asset?\n\n### Response:\nSubstandard Asset: NPA for less than 12 months"
    },
    {
        "instruction": "What is Prime numbers?",
        "input": "",
        "output": "Prime numbers: divisible only by 1 and itself. Example: 2, 3, 5, 7, 11, 13",
        "text": "### Instruction:\nWhat is Prime numbers?\n\n### Response:\nPrime numbers: divisible only by 1 and itself. Example: 2, 3, 5, 7, 11, 13"
    },
    {
        "instruction": "What is the formula for --- Evaluation Metrics ---?",
        "input": "",
        "output": "- Accuracy = (TP + TN) / (TP + TN + FP + FN)",
        "text": "### Instruction:\nWhat is the formula for --- Evaluation Metrics ---?\n\n### Response:\n- Accuracy = (TP + TN) / (TP + TN + FP + FN)"
    },
    {
        "instruction": "What is Storage hierarchy?",
        "input": "",
        "output": "Storage hierarchy: Register > Cache > RAM > Hard Disk",
        "text": "### Instruction:\nWhat is Storage hierarchy?\n\n### Response:\nStorage hierarchy: Register > Cache > RAM > Hard Disk"
    },
    {
        "instruction": "What is CIBIL Score?",
        "input": "",
        "output": "CIBIL Score: credit score ranging 300-900 (750+ considered good)",
        "text": "### Instruction:\nWhat is CIBIL Score?\n\n### Response:\nCIBIL Score: credit score ranging 300-900 (750+ considered good)"
    },
    {
        "instruction": "What is the formula for P(A or B) = P(A) + P(B) - P(A and B)?",
        "input": "",
        "output": "Combination nCr = n! / (r! x (n-r)!)",
        "text": "### Instruction:\nWhat is the formula for P(A or B) = P(A) + P(B) - P(A and B)?\n\n### Response:\nCombination nCr = n! / (r! x (n-r)!)"
    },
    {
        "instruction": "What is Batch Normalization?",
        "input": "",
        "output": "Batch Normalization: normalizes layer inputs, faster training, acts as regularizer",
        "text": "### Instruction:\nWhat is Batch Normalization?\n\n### Response:\nBatch Normalization: normalizes layer inputs, faster training, acts as regularizer"
    },
    {
        "instruction": "What is Tech?",
        "input": "",
        "output": "Tech: Python, Streamlit, Supabase, OpenAI Whisper, HuggingFace Spaces",
        "text": "### Instruction:\nWhat is Tech?\n\n### Response:\nTech: Python, Streamlit, Supabase, OpenAI Whisper, HuggingFace Spaces"
    },
    {
        "instruction": "What is the formula for When compounded half-yearly: Rate = R/2, Time = 2T?",
        "input": "",
        "output": "Difference between CI and SI for 3 years = P x (R/100)^2 x (R/100 + 3)",
        "text": "### Instruction:\nWhat is the formula for When compounded half-yearly: Rate = R/2, Time = 2T?\n\n### Response:\nDifference between CI and SI for 3 years = P x (R/100)^2 x (R/100 + 3)"
    },
    {
        "instruction": "What is Cooperative Banks?",
        "input": "",
        "output": "Cooperative Banks: serve specific communities",
        "text": "### Instruction:\nWhat is Cooperative Banks?\n\n### Response:\nCooperative Banks: serve specific communities"
    },
    {
        "instruction": "What is the formula for Mother's father = Maternal grandfather?",
        "input": "",
        "output": "Son's wife = Daughter-in-law",
        "text": "### Instruction:\nWhat is the formula for Mother's father = Maternal grandfather?\n\n### Response:\nSon's wife = Daughter-in-law"
    },
    {
        "instruction": "What is the formula for Average = Sum of observations / Number of observations?",
        "input": "",
        "output": "New average = A + (y-x)/n",
        "text": "### Instruction:\nWhat is the formula for Average = Sum of observations / Number of observations?\n\n### Response:\nNew average = A + (y-x)/n"
    },
    {
        "instruction": "What is Western Ghats?",
        "input": "",
        "output": "Western Ghats: runs parallel to western coast, biodiversity hotspot",
        "text": "### Instruction:\nWhat is Western Ghats?\n\n### Response:\nWestern Ghats: runs parallel to western coast, biodiversity hotspot"
    },
    {
        "instruction": "What is Feature Maps?",
        "input": "",
        "output": "Feature Maps: output of convolutional layer",
        "text": "### Instruction:\nWhat is Feature Maps?\n\n### Response:\nFeature Maps: output of convolutional layer"
    },
    {
        "instruction": "What is Cash Reserve Ratio (CRR)?",
        "input": "",
        "output": "Cash Reserve Ratio (CRR): percentage of deposits banks must keep with RBI",
        "text": "### Instruction:\nWhat is Cash Reserve Ratio (CRR)?\n\n### Response:\nCash Reserve Ratio (CRR): percentage of deposits banks must keep with RBI"
    },
    {
        "instruction": "What are the key concepts in Number System?",
        "input": "",
        "output": "Natural numbers: 1, 2, 3, 4, ...\nWhole numbers: 0, 1, 2, 3, ...\nIntegers: ...-3, -2, -1, 0, 1, 2, 3...\nRational numbers: can be expressed as p/q where q is not zero\nIrrational numbers: cannot be expressed as p/q. Example: sqrt(2), pi\nPrime numbers: divisible only by 1 and itself. Example: 2, 3, 5, 7, 11, 13\nComposite numbers: more than 2 factors. Example: 4, 6, 8, 9\nHCF (Highest Common Factor): largest number dividing two numbers exactly\nLCM (Least Common Multiple): smallest number divisible by both numbers\nHCF x LCM = Product of two numbers\nDivisibility by 2: last digit is 0, 2, 4, 6, or 8\nDi",
        "text": "### Instruction:\nWhat are the key concepts in Number System?\n\n### Response:\nNatural numbers: 1, 2, 3, 4, ...\nWhole numbers: 0, 1, 2, 3, ...\nIntegers: ...-3, -2, -1, 0, 1, 2, 3...\nRational numbers: can be expressed as p/q where q is not zero\nIrrational numbers: cannot be expressed as p/q. Example: sqrt(2), pi\nPrime numbers: divisible only by 1 and itself. Example: 2, 3, 5, 7, 11, 13\nComposite numbers: more than 2 factors. Example: 4, 6, 8, 9\nHCF (Highest Common Factor): largest number dividing two numbers exactly\nLCM (Least Common Multiple): smallest number divisible by both numbers\nHCF x LCM = Product of two numbers\nDivisibility by 2: last digit is 0, 2, 4, 6, or 8\nDi"
    },
    {
        "instruction": "What is the significance/value of NITI Aayog?",
        "input": "",
        "output": "NITI Aayog: replaced Planning Commission in 2015",
        "text": "### Instruction:\nWhat is the significance/value of NITI Aayog?\n\n### Response:\nNITI Aayog: replaced Planning Commission in 2015"
    },
    {
        "instruction": "What is Savings Account?",
        "input": "",
        "output": "Savings Account: for individuals, interest paid, transaction limits",
        "text": "### Instruction:\nWhat is Savings Account?\n\n### Response:\nSavings Account: for individuals, interest paid, transaction limits"
    },
    {
        "instruction": "What is Article 20?",
        "input": "",
        "output": "Article 20: Protection against self-incrimination",
        "text": "### Instruction:\nWhat is Article 20?\n\n### Response:\nArticle 20: Protection against self-incrimination"
    },
    {
        "instruction": "What is Article 22?",
        "input": "",
        "output": "Article 22: Protection against arbitrary arrest",
        "text": "### Instruction:\nWhat is Article 22?\n\n### Response:\nArticle 22: Protection against arbitrary arrest"
    },
    {
        "instruction": "What is Article 16?",
        "input": "",
        "output": "Article 16: Equality of opportunity in public employment",
        "text": "### Instruction:\nWhat is Article 16?\n\n### Response:\nArticle 16: Equality of opportunity in public employment"
    },
    {
        "instruction": "What is Mirror image?",
        "input": "",
        "output": "Mirror image: A corresponds to Z, B to Y, C to X, etc.",
        "text": "### Instruction:\nWhat is Mirror image?\n\n### Response:\nMirror image: A corresponds to Z, B to Y, C to X, etc."
    },
    {
        "instruction": "What bugs did you find in Drive-Awake and how would you fix them?",
        "input": "",
        "output": "Three main bugs:\n1. Duplicate packages in requirements.txt (tensorflow, numpy, opencv, pygame)\n   Fix: remove duplicates, keep single version of each\n2. Dockerfile uses dnf (RHEL) but base image is python:3.10-slim (Debian)\n   Fix: use apt-get instead of dnf, or change base to fedora image\n3. float(left_pred) fails on 2D softmax array\n   Fix: use left_pred[1] or left_pred[0][1] to get closed-eye probability",
        "text": "### Instruction:\nWhat bugs did you find in Drive-Awake and how would you fix them?\n\n### Response:\nThree main bugs:\n1. Duplicate packages in requirements.txt (tensorflow, numpy, opencv, pygame)\n   Fix: remove duplicates, keep single version of each\n2. Dockerfile uses dnf (RHEL) but base image is python:3.10-slim (Debian)\n   Fix: use apt-get instead of dnf, or change base to fedora image\n3. float(left_pred) fails on 2D softmax array\n   Fix: use left_pred[1] or left_pred[0][1] to get closed-eye probability"
    },
    {
        "instruction": "What is the formula for --- Mensuration ---?",
        "input": "",
        "output": "Square: Area = side^2, Perimeter = 4 x side",
        "text": "### Instruction:\nWhat is the formula for --- Mensuration ---?\n\n### Response:\nSquare: Area = side^2, Perimeter = 4 x side"
    },
    {
        "instruction": "What is it for a new task. Feature extraction?",
        "input": "",
        "output": "it for a new task. Feature extraction: freeze pretrained layers, train new",
        "text": "### Instruction:\nWhat is it for a new task. Feature extraction?\n\n### Response:\nit for a new task. Feature extraction: freeze pretrained layers, train new"
    },
    {
        "instruction": "What are the key concepts in Time, Speed and Distance?",
        "input": "",
        "output": "Speed = Distance / Time\nDistance = Speed x Time\nTime = Distance / Speed\nAverage Speed = Total Distance / Total Time\nIf two trains of length L1 and L2 cross each other:\n  - Moving same direction: Time = (L1+L2) / (S1-S2)\n  - Moving opposite direction: Time = (L1+L2) / (S1+S2)\n1 km/h = 5/18 m/s\n1 m/s = 18/5 km/h\nRelative speed (same direction) = S1 - S2\nRelative speed (opposite direction) = S1 + S2",
        "text": "### Instruction:\nWhat are the key concepts in Time, Speed and Distance?\n\n### Response:\nSpeed = Distance / Time\nDistance = Speed x Time\nTime = Distance / Speed\nAverage Speed = Total Distance / Total Time\nIf two trains of length L1 and L2 cross each other:\n  - Moving same direction: Time = (L1+L2) / (S1-S2)\n  - Moving opposite direction: Time = (L1+L2) / (S1+S2)\n1 km/h = 5/18 m/s\n1 m/s = 18/5 km/h\nRelative speed (same direction) = S1 - S2\nRelative speed (opposite direction) = S1 + S2"
    },
    {
        "instruction": "What is Eastern Ghats?",
        "input": "",
        "output": "Eastern Ghats: discontinuous range along eastern coast",
        "text": "### Instruction:\nWhat is Eastern Ghats?\n\n### Response:\nEastern Ghats: discontinuous range along eastern coast"
    },
    {
        "instruction": "What is Linear Regression?",
        "input": "",
        "output": "Linear Regression: predict continuous values, fit line to data",
        "text": "### Instruction:\nWhat is Linear Regression?\n\n### Response:\nLinear Regression: predict continuous values, fit line to data"
    },
    {
        "instruction": "What is Tokenization?",
        "input": "",
        "output": "Tokenization: splitting text into tokens (words, subwords, characters)",
        "text": "### Instruction:\nWhat is Tokenization?\n\n### Response:\nTokenization: splitting text into tokens (words, subwords, characters)"
    },
    {
        "instruction": "What is Whole numbers?",
        "input": "",
        "output": "Whole numbers: 0, 1, 2, 3, ...",
        "text": "### Instruction:\nWhat is Whole numbers?\n\n### Response:\nWhole numbers: 0, 1, 2, 3, ..."
    },
    {
        "instruction": "What is RBI regulates?",
        "input": "",
        "output": "RBI regulates: commercial banks, cooperative banks, NBFCs",
        "text": "### Instruction:\nWhat is RBI regulates?\n\n### Response:\nRBI regulates: commercial banks, cooperative banks, NBFCs"
    },
    {
        "instruction": "What is the significance/value of Article 360?",
        "input": "",
        "output": "Article 360: Financial Emergency",
        "text": "### Instruction:\nWhat is the significance/value of Article 360?\n\n### Response:\nArticle 360: Financial Emergency"
    },
    {
        "instruction": "What is Cone?",
        "input": "",
        "output": "Cone: Volume = 1/3 x pi x r^2 x h, Slant height = sqrt(r^2 + h^2)",
        "text": "### Instruction:\nWhat is Cone?\n\n### Response:\nCone: Volume = 1/3 x pi x r^2 x h, Slant height = sqrt(r^2 + h^2)"
    },
    {
        "instruction": "What is the formula for - Moving same direction: Time = (L1+L2) / (S1-S2)?",
        "input": "",
        "output": "1 m/s = 18/5 km/h",
        "text": "### Instruction:\nWhat is the formula for - Moving same direction: Time = (L1+L2) / (S1-S2)?\n\n### Response:\n1 m/s = 18/5 km/h"
    },
    {
        "instruction": "What is Pooling Layer?",
        "input": "",
        "output": "Pooling Layer: reduces spatial dimensions (Max Pooling, Average Pooling)",
        "text": "### Instruction:\nWhat is Pooling Layer?\n\n### Response:\nPooling Layer: reduces spatial dimensions (Max Pooling, Average Pooling)"
    },
    {
        "instruction": "What is 2 October?",
        "input": "",
        "output": "2 October: Gandhi Jayanti",
        "text": "### Instruction:\nWhat is 2 October?\n\n### Response:\n2 October: Gandhi Jayanti"
    },
    {
        "instruction": "What is SECTION 1?",
        "input": "",
        "output": "SECTION 1: QUANTITATIVE APTITUDE",
        "text": "### Instruction:\nWhat is SECTION 1?\n\n### Response:\nSECTION 1: QUANTITATIVE APTITUDE"
    },
    {
        "instruction": "What is If a?",
        "input": "",
        "output": "If a: b = c:d then ad = bc (cross multiplication)",
        "text": "### Instruction:\nWhat is If a?\n\n### Response:\nIf a: b = c:d then ad = bc (cross multiplication)"
    },
    {
        "instruction": "What is Demand Draft?",
        "input": "",
        "output": "Demand Draft: guaranteed payment instrument",
        "text": "### Instruction:\nWhat is Demand Draft?\n\n### Response:\nDemand Draft: guaranteed payment instrument"
    },
    {
        "instruction": "What is BERT?",
        "input": "",
        "output": "BERT: Bidirectional Encoder, pre-trained on masked language modeling",
        "text": "### Instruction:\nWhat is BERT?\n\n### Response:\nBERT: Bidirectional Encoder, pre-trained on masked language modeling"
    },
    {
        "instruction": "What is Stride?",
        "input": "",
        "output": "Stride: step size of filter movement",
        "text": "### Instruction:\nWhat is Stride?\n\n### Response:\nStride: step size of filter movement"
    },
    {
        "instruction": "What is Article 112?",
        "input": "",
        "output": "Article 112: Annual Financial Statement (Union Budget)",
        "text": "### Instruction:\nWhat is Article 112?\n\n### Response:\nArticle 112: Annual Financial Statement (Union Budget)"
    },
    {
        "instruction": "What is DBMS?",
        "input": "",
        "output": "DBMS: Database Management System (MySQL, Oracle, PostgreSQL)",
        "text": "### Instruction:\nWhat is DBMS?\n\n### Response:\nDBMS: Database Management System (MySQL, Oracle, PostgreSQL)"
    },
    {
        "instruction": "What is the formula for For mutually exclusive events: P(A or B) = P(A) + P(B)?",
        "input": "",
        "output": "Permutation nPr = n! / (n-r)!",
        "text": "### Instruction:\nWhat is the formula for For mutually exclusive events: P(A or B) = P(A) + P(B)?\n\n### Response:\nPermutation nPr = n! / (n-r)!"
    },
    {
        "instruction": "What is Fully Connected Layer?",
        "input": "",
        "output": "Fully Connected Layer: final classification layer",
        "text": "### Instruction:\nWhat is Fully Connected Layer?\n\n### Response:\nFully Connected Layer: final classification layer"
    },
    {
        "instruction": "What is the significance/value of RBI established?",
        "input": "",
        "output": "RBI established: 1 April 1935",
        "text": "### Instruction:\nWhat is the significance/value of RBI established?\n\n### Response:\nRBI established: 1 April 1935"
    },
    {
        "instruction": "What is the formula for SECTION 8: PARTH'S PROJECT SUMMARIES (FOR INTERVIEWS)?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for SECTION 8: PARTH'S PROJECT SUMMARIES (FOR INTERVIEWS)?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "A and B can complete a work in 12 days and 18 days respectively. In how many days can they complete it together?",
        "output": "A's 1 day work = 1/12, B's 1 day work = 1/18. Together = 1/12 + 1/18 = 3/36 + 2/36 = 5/36. Days = 36/5 = 7.2 days",
        "text": "### Instruction:\nA and B can complete a work in 12 days and 18 days respectively. In how many days can they complete it together?\n\n### Response:\nA's 1 day work = 1/12, B's 1 day work = 1/18. Together = 1/12 + 1/18 = 3/36 + 2/36 = 5/36. Days = 36/5 = 7.2 days"
    },
    {
        "instruction": "What is the significance/value of Article 19?",
        "input": "",
        "output": "Article 19: Freedom of speech, assembly, movement, profession",
        "text": "### Instruction:\nWhat is the significance/value of Article 19?\n\n### Response:\nArticle 19: Freedom of speech, assembly, movement, profession"
    },
    {
        "instruction": "What is HCF (Highest Common Factor)?",
        "input": "",
        "output": "HCF (Highest Common Factor): largest number dividing two numbers exactly",
        "text": "### Instruction:\nWhat is HCF (Highest Common Factor)?\n\n### Response:\nHCF (Highest Common Factor): largest number dividing two numbers exactly"
    },
    {
        "instruction": "What is the formula for Weighted average = (w1*x1 + w2*x2 + ...) / (w1 + w2 + ...)?",
        "input": "",
        "output": "Rectangle: Area = length x breadth, Perimeter = 2(l+b)",
        "text": "### Instruction:\nWhat is the formula for Weighted average = (w1*x1 + w2*x2 + ...) / (w1 + w2 + ...)?\n\n### Response:\nRectangle: Area = length x breadth, Perimeter = 2(l+b)"
    },
    {
        "instruction": "What is Conjunctions?",
        "input": "",
        "output": "Conjunctions: and, but, or, nor, for, yet, so (FANBOYS)",
        "text": "### Instruction:\nWhat is Conjunctions?\n\n### Response:\nConjunctions: and, but, or, nor, for, yet, so (FANBOYS)"
    },
    {
        "instruction": "What is Pradhan Mantri Jeevan Jyoti Bima Yojana (PMJJBY)?",
        "input": "",
        "output": "Pradhan Mantri Jeevan Jyoti Bima Yojana (PMJJBY): life insurance scheme",
        "text": "### Instruction:\nWhat is Pradhan Mantri Jeevan Jyoti Bima Yojana (PMJJBY)?\n\n### Response:\nPradhan Mantri Jeevan Jyoti Bima Yojana (PMJJBY): life insurance scheme"
    },
    {
        "instruction": "What is TensorFlow?",
        "input": "",
        "output": "TensorFlow: model.save(), tf.keras.models.load_model()",
        "text": "### Instruction:\nWhat is TensorFlow?\n\n### Response:\nTensorFlow: model.save(), tf.keras.models.load_model()"
    },
    {
        "instruction": "What is SQL?",
        "input": "",
        "output": "SQL: Structured Query Language",
        "text": "### Instruction:\nWhat is SQL?\n\n### Response:\nSQL: Structured Query Language"
    },
    {
        "instruction": "What is Layer normalization?",
        "input": "",
        "output": "Layer normalization: stabilizes training",
        "text": "### Instruction:\nWhat is Layer normalization?\n\n### Response:\nLayer normalization: stabilizes training"
    },
    {
        "instruction": "What is Article 356?",
        "input": "",
        "output": "Article 356: President's Rule in states",
        "text": "### Instruction:\nWhat is Article 356?\n\n### Response:\nArticle 356: President's Rule in states"
    },
    {
        "instruction": "What is the formula for - Always lower than repo rate?",
        "input": "",
        "output": "- Increase in CRR = less money with banks = less credit = inflation control",
        "text": "### Instruction:\nWhat is the formula for - Always lower than repo rate?\n\n### Response:\n- Increase in CRR = less money with banks = less credit = inflation control"
    },
    {
        "instruction": "What is Startup India?",
        "input": "",
        "output": "Startup India: support for startups and entrepreneurs",
        "text": "### Instruction:\nWhat is Startup India?\n\n### Response:\nStartup India: support for startups and entrepreneurs"
    },
    {
        "instruction": "What is Register?",
        "input": "",
        "output": "Register: fastest, inside CPU",
        "text": "### Instruction:\nWhat is Register?\n\n### Response:\nRegister: fastest, inside CPU"
    },
    {
        "instruction": "What is India's inflation measure?",
        "input": "",
        "output": "India's inflation measure: CPI (Consumer Price Index) - primary",
        "text": "### Instruction:\nWhat is India's inflation measure?\n\n### Response:\nIndia's inflation measure: CPI (Consumer Price Index) - primary"
    },
    {
        "instruction": "What is the significance/value of Article 15?",
        "input": "",
        "output": "Article 15: Prohibition of discrimination",
        "text": "### Instruction:\nWhat is the significance/value of Article 15?\n\n### Response:\nArticle 15: Prohibition of discrimination"
    },
    {
        "instruction": "What is ResNet?",
        "input": "",
        "output": "ResNet: residual connections, 150+ layers, solves vanishing gradient",
        "text": "### Instruction:\nWhat is ResNet?\n\n### Response:\nResNet: residual connections, 150+ layers, solves vanishing gradient"
    },
    {
        "instruction": "What is Check grammar?",
        "input": "",
        "output": "Check grammar: noun/verb/adjective/adverb required",
        "text": "### Instruction:\nWhat is Check grammar?\n\n### Response:\nCheck grammar: noun/verb/adjective/adverb required"
    },
    {
        "instruction": "What is Solution?",
        "input": "",
        "output": "Solution: more data, regularization, dropout, early stopping",
        "text": "### Instruction:\nWhat is Solution?\n\n### Response:\nSolution: more data, regularization, dropout, early stopping"
    },
    {
        "instruction": "What is the formula for Stand-Up India: loans to SC/ST and women entrepreneurs?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for Stand-Up India: loans to SC/ST and women entrepreneurs?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What are the key concepts in Reserve Bank of India (RBI)?",
        "input": "",
        "output": "RBI established: 1 April 1935\nRBI nationalized: 1 January 1949\nRBI headquarters: Mumbai\nRBI Governor: appointed by Central Government\nRBI is banker to government and banker of banks\nFunctions: monetary policy, currency issue, foreign exchange management\nRBI regulates: commercial banks, cooperative banks, NBFCs\nLender of last resort: RBI provides emergency funds to banks",
        "text": "### Instruction:\nWhat are the key concepts in Reserve Bank of India (RBI)?\n\n### Response:\nRBI established: 1 April 1935\nRBI nationalized: 1 January 1949\nRBI headquarters: Mumbai\nRBI Governor: appointed by Central Government\nRBI is banker to government and banker of banks\nFunctions: monetary policy, currency issue, foreign exchange management\nRBI regulates: commercial banks, cooperative banks, NBFCs\nLender of last resort: RBI provides emergency funds to banks"
    },
    {
        "instruction": "What is Image preprocessing?",
        "input": "",
        "output": "Image preprocessing: resize, normalize, augment",
        "text": "### Instruction:\nWhat is Image preprocessing?\n\n### Response:\nImage preprocessing: resize, normalize, augment"
    },
    {
        "instruction": "What is SECTION 7?",
        "input": "",
        "output": "SECTION 7: AI/ML INTERVIEW PREPARATION",
        "text": "### Instruction:\nWhat is SECTION 7?\n\n### Response:\nSECTION 7: AI/ML INTERVIEW PREPARATION"
    },
    {
        "instruction": "What are the key concepts in Profit and Loss?",
        "input": "",
        "output": "Profit = Selling Price - Cost Price\nLoss = Cost Price - Selling Price\nProfit% = (Profit / Cost Price) x 100\nLoss% = (Loss / Cost Price) x 100\nSelling Price = Cost Price x (1 + Profit%/100)\nCost Price = Selling Price / (1 + Profit%/100)\nIf two items sold at same price, one at x% profit and one at x% loss, overall loss% = x^2/100\nMarked Price: price written on item before discount\nDiscount = Marked Price - Selling Price\nDiscount% = (Discount / Marked Price) x 100",
        "text": "### Instruction:\nWhat are the key concepts in Profit and Loss?\n\n### Response:\nProfit = Selling Price - Cost Price\nLoss = Cost Price - Selling Price\nProfit% = (Profit / Cost Price) x 100\nLoss% = (Loss / Cost Price) x 100\nSelling Price = Cost Price x (1 + Profit%/100)\nCost Price = Selling Price / (1 + Profit%/100)\nIf two items sold at same price, one at x% profit and one at x% loss, overall loss% = x^2/100\nMarked Price: price written on item before discount\nDiscount = Marked Price - Selling Price\nDiscount% = (Discount / Marked Price) x 100"
    },
    {
        "instruction": "What is Cuboid?",
        "input": "",
        "output": "Cuboid: Volume = l x b x h, Surface Area = 2(lb + bh + lh)",
        "text": "### Instruction:\nWhat is Cuboid?\n\n### Response:\nCuboid: Volume = l x b x h, Surface Area = 2(lb + bh + lh)"
    },
    {
        "instruction": "What is Scikit-learn?",
        "input": "",
        "output": "Scikit-learn: ML algorithms, preprocessing, evaluation",
        "text": "### Instruction:\nWhat is Scikit-learn?\n\n### Response:\nScikit-learn: ML algorithms, preprocessing, evaluation"
    },
    {
        "instruction": "What is the formula for - Used in: Music Recommender project (Parth's project)?",
        "input": "",
        "output": "- Based on Bayes theorem: P(A|B) = P(B|A) x P(A) / P(B)",
        "text": "### Instruction:\nWhat is the formula for - Used in: Music Recommender project (Parth's project)?\n\n### Response:\n- Based on Bayes theorem: P(A|B) = P(B|A) x P(A) / P(B)"
    },
    {
        "instruction": "What is the formula for Combination nCr = n! / (r! x (n-r)!)?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for Combination nCr = n! / (r! x (n-r)!)?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is MAE (Mean Absolute Error)?",
        "input": "",
        "output": "MAE (Mean Absolute Error): average absolute difference",
        "text": "### Instruction:\nWhat is MAE (Mean Absolute Error)?\n\n### Response:\nMAE (Mean Absolute Error): average absolute difference"
    },
    {
        "instruction": "What is the formula for --- Direction and Distance ---?",
        "input": "",
        "output": "Turning right from North = East",
        "text": "### Instruction:\nWhat is the formula for --- Direction and Distance ---?\n\n### Response:\nTurning right from North = East"
    },
    {
        "instruction": "What is Irrational numbers?",
        "input": "",
        "output": "Irrational numbers: cannot be expressed as p/q. Example: sqrt(2), pi",
        "text": "### Instruction:\nWhat is Irrational numbers?\n\n### Response:\nIrrational numbers: cannot be expressed as p/q. Example: sqrt(2), pi"
    },
    {
        "instruction": "What is Feature importance?",
        "input": "",
        "output": "Feature importance: how much each feature contributes",
        "text": "### Instruction:\nWhat is Feature importance?\n\n### Response:\nFeature importance: how much each feature contributes"
    },
    {
        "instruction": "What is Multi-head attention?",
        "input": "",
        "output": "Multi-head attention: run attention multiple times in parallel",
        "text": "### Instruction:\nWhat is Multi-head attention?\n\n### Response:\nMulti-head attention: run attention multiple times in parallel"
    },
    {
        "instruction": "What are the key concepts in Error Spotting?",
        "input": "",
        "output": "Check subject-verb agreement\nCheck pronoun-antecedent agreement\nCheck tense consistency\nCheck article usage\nCheck preposition usage\nCheck parallel structure in lists",
        "text": "### Instruction:\nWhat are the key concepts in Error Spotting?\n\n### Response:\nCheck subject-verb agreement\nCheck pronoun-antecedent agreement\nCheck tense consistency\nCheck article usage\nCheck preposition usage\nCheck parallel structure in lists"
    },
    {
        "instruction": "What is the significance/value of India's GDP growth rate?",
        "input": "",
        "output": "India's GDP growth rate: around 6-7% annually",
        "text": "### Instruction:\nWhat is the significance/value of India's GDP growth rate?\n\n### Response:\nIndia's GDP growth rate: around 6-7% annually"
    },
    {
        "instruction": "What is the formula for Father's brother = Uncle, Father's sister = Aunt?",
        "input": "",
        "output": "Brother's son = Nephew, Brother's daughter = Niece",
        "text": "### Instruction:\nWhat is the formula for Father's brother = Uncle, Father's sister = Aunt?\n\n### Response:\nBrother's son = Nephew, Brother's daughter = Niece"
    },
    {
        "instruction": "What is Cube series?",
        "input": "",
        "output": "Cube series: 1,8,27,64,125...",
        "text": "### Instruction:\nWhat is Cube series?\n\n### Response:\nCube series: 1,8,27,64,125..."
    },
    {
        "instruction": "What is Bugs to fix?",
        "input": "",
        "output": "Bugs to fix: duplicate requirements, Dockerfile OS mismatch, float() on array",
        "text": "### Instruction:\nWhat is Bugs to fix?\n\n### Response:\nBugs to fix: duplicate requirements, Dockerfile OS mismatch, float() on array"
    },
    {
        "instruction": "What is Divisibility by 3?",
        "input": "",
        "output": "Divisibility by 3: sum of digits divisible by 3",
        "text": "### Instruction:\nWhat is Divisibility by 3?\n\n### Response:\nDivisibility by 3: sum of digits divisible by 3"
    },
    {
        "instruction": "What is the formula for Cylinder: Volume = pi x r^2 x h, Curved Surface Area = 2 x pi x r x h?",
        "input": "",
        "output": "Cuboid: Volume = l x b x h, Surface Area = 2(lb + bh + lh)",
        "text": "### Instruction:\nWhat is the formula for Cylinder: Volume = pi x r^2 x h, Curved Surface Area = 2 x pi x r x h?\n\n### Response:\nCuboid: Volume = l x b x h, Surface Area = 2(lb + bh + lh)"
    },
    {
        "instruction": "What is 14 April?",
        "input": "",
        "output": "14 April: Ambedkar Jayanti",
        "text": "### Instruction:\nWhat is 14 April?\n\n### Response:\n14 April: Ambedkar Jayanti"
    },
    {
        "instruction": "What is RBI nationalized?",
        "input": "",
        "output": "RBI nationalized: 1 January 1949",
        "text": "### Instruction:\nWhat is RBI nationalized?\n\n### Response:\nRBI nationalized: 1 January 1949"
    },
    {
        "instruction": "What is Lender of last resort?",
        "input": "",
        "output": "Lender of last resort: RBI provides emergency funds to banks",
        "text": "### Instruction:\nWhat is Lender of last resort?\n\n### Response:\nLender of last resort: RBI provides emergency funds to banks"
    },
    {
        "instruction": "What is the formula for - cv2.resize(): resize image?",
        "input": "",
        "output": "- EAR < 0.25 for consecutive frames = drowsiness detected",
        "text": "### Instruction:\nWhat is the formula for - cv2.resize(): resize image?\n\n### Response:\n- EAR < 0.25 for consecutive frames = drowsiness detected"
    },
    {
        "instruction": "What is POP3/IMAP?",
        "input": "",
        "output": "POP3/IMAP: receiving emails",
        "text": "### Instruction:\nWhat is POP3/IMAP?\n\n### Response:\nPOP3/IMAP: receiving emails"
    },
    {
        "instruction": "What is Transfer Learning?",
        "input": "",
        "output": "Transfer Learning: use pre-trained model weights on new task",
        "text": "### Instruction:\nWhat is Transfer Learning?\n\n### Response:\nTransfer Learning: use pre-trained model weights on new task"
    },
    {
        "instruction": "What is the formula for --- Series Completion ---?",
        "input": "",
        "output": "Fibonacci: each term = sum of previous two (1,1,2,3,5,8...)",
        "text": "### Instruction:\nWhat is the formula for --- Series Completion ---?\n\n### Response:\nFibonacci: each term = sum of previous two (1,1,2,3,5,8...)"
    },
    {
        "instruction": "What is Check meaning?",
        "input": "",
        "output": "Check meaning: word must fit context",
        "text": "### Instruction:\nWhat is Check meaning?\n\n### Response:\nCheck meaning: word must fit context"
    },
    {
        "instruction": "The ratio of A to B is 3:4 and B to C is 2:3. Find A:B:C.",
        "output": "A:B = 3:4, B:C = 2:3. Multiply to get common B: A:B = 6:8, B:C = 8:12. So A:B:C = 6:8:12 = 3:4:6",
        "text": "### Instruction:\nThe ratio of A to B is 3:4 and B to C is 2:3. Find A:B:C.\n\n### Response:\nA:B = 3:4, B:C = 2:3. Multiply to get common B: A:B = 6:8, B:C = 8:12. So A:B:C = 6:8:12 = 3:4:6"
    },
    {
        "instruction": "What are the key concepts in Series Completion?",
        "input": "",
        "output": "Arithmetic series: difference is constant (2,4,6,8...)\nGeometric series: ratio is constant (2,4,8,16...)\nFibonacci: each term = sum of previous two (1,1,2,3,5,8...)\nSquare series: 1,4,9,16,25...\nCube series: 1,8,27,64,125...\nMixed series: combine two or more patterns",
        "text": "### Instruction:\nWhat are the key concepts in Series Completion?\n\n### Response:\nArithmetic series: difference is constant (2,4,6,8...)\nGeometric series: ratio is constant (2,4,8,16...)\nFibonacci: each term = sum of previous two (1,1,2,3,5,8...)\nSquare series: 1,4,9,16,25...\nCube series: 1,8,27,64,125...\nMixed series: combine two or more patterns"
    },
    {
        "instruction": "What is Pradhan Mantri Jan Dhan Yojana (PMJDY)?",
        "input": "",
        "output": "Pradhan Mantri Jan Dhan Yojana (PMJDY): financial inclusion scheme",
        "text": "### Instruction:\nWhat is Pradhan Mantri Jan Dhan Yojana (PMJDY)?\n\n### Response:\nPradhan Mantri Jan Dhan Yojana (PMJDY): financial inclusion scheme"
    },
    {
        "instruction": "What are the key concepts in Database?",
        "input": "",
        "output": "Database: organized collection of data\nDBMS: Database Management System (MySQL, Oracle, PostgreSQL)\nSQL: Structured Query Language\nSELECT: retrieve data\nINSERT: add data\nUPDATE: modify data\nDELETE: remove data\nPrimary Key: unique identifier for each record\nForeign Key: links two tables",
        "text": "### Instruction:\nWhat are the key concepts in Database?\n\n### Response:\nDatabase: organized collection of data\nDBMS: Database Management System (MySQL, Oracle, PostgreSQL)\nSQL: Structured Query Language\nSELECT: retrieve data\nINSERT: add data\nUPDATE: modify data\nDELETE: remove data\nPrimary Key: unique identifier for each record\nForeign Key: links two tables"
    },
    {
        "instruction": "What is the significance/value of Article 16?",
        "input": "",
        "output": "Article 16: Equality of opportunity in public employment",
        "text": "### Instruction:\nWhat is the significance/value of Article 16?\n\n### Response:\nArticle 16: Equality of opportunity in public employment"
    },
    {
        "instruction": "What is Bank Rate?",
        "input": "",
        "output": "Bank Rate: rate at which RBI provides long-term loans to banks",
        "text": "### Instruction:\nWhat is Bank Rate?\n\n### Response:\nBank Rate: rate at which RBI provides long-term loans to banks"
    },
    {
        "instruction": "What is NPA (Non-Performing Asset)?",
        "input": "",
        "output": "NPA (Non-Performing Asset): loan where interest/principal overdue 90+ days",
        "text": "### Instruction:\nWhat is NPA (Non-Performing Asset)?\n\n### Response:\nNPA (Non-Performing Asset): loan where interest/principal overdue 90+ days"
    },
    {
        "instruction": "What is the formula for Time = Distance / Speed?",
        "input": "",
        "output": "- Moving same direction: Time = (L1+L2) / (S1-S2)",
        "text": "### Instruction:\nWhat is the formula for Time = Distance / Speed?\n\n### Response:\n- Moving same direction: Time = (L1+L2) / (S1-S2)"
    },
    {
        "instruction": "What is Attention formula?",
        "input": "",
        "output": "Attention formula: softmax(QK^T / sqrt(d_k)) x V",
        "text": "### Instruction:\nWhat is Attention formula?\n\n### Response:\nAttention formula: softmax(QK^T / sqrt(d_k)) x V"
    },
    {
        "instruction": "What is Train/Validation/Test split?",
        "input": "",
        "output": "Train/Validation/Test split: typically 70/15/15 or 80/10/10",
        "text": "### Instruction:\nWhat is Train/Validation/Test split?\n\n### Response:\nTrain/Validation/Test split: typically 70/15/15 or 80/10/10"
    },
    {
        "instruction": "What is Dropout?",
        "input": "",
        "output": "Dropout: randomly deactivates neurons during training, prevents overfitting",
        "text": "### Instruction:\nWhat is Dropout?\n\n### Response:\nDropout: randomly deactivates neurons during training, prevents overfitting"
    },
    {
        "instruction": "What is L1 (Lasso)?",
        "input": "",
        "output": "L1 (Lasso): adds absolute value of weights to loss, produces sparse models",
        "text": "### Instruction:\nWhat is L1 (Lasso)?\n\n### Response:\nL1 (Lasso): adds absolute value of weights to loss, produces sparse models"
    },
    {
        "instruction": "What is the significance/value of Premium?",
        "input": "",
        "output": "Premium: 436 per year, Cover: 2 lakh",
        "text": "### Instruction:\nWhat is the significance/value of Premium?\n\n### Response:\nPremium: 436 per year, Cover: 2 lakh"
    },
    {
        "instruction": "Explain the vanishing gradient problem.",
        "input": "",
        "output": "During backpropagation in deep networks, gradients become very small\nas they flow backward through many layers. Activations like sigmoid squash\nvalues to 0-1, making gradients tiny. Solution: ReLU activation, residual\nconnections (ResNet), LSTM for sequences, batch normalization.",
        "text": "### Instruction:\nExplain the vanishing gradient problem.\n\n### Response:\nDuring backpropagation in deep networks, gradients become very small\nas they flow backward through many layers. Activations like sigmoid squash\nvalues to 0-1, making gradients tiny. Solution: ReLU activation, residual\nconnections (ResNet), LSTM for sequences, batch normalization."
    },
    {
        "instruction": "What is the formula for Encryption: converting data to unreadable format?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for Encryption: converting data to unreadable format?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is Tenses?",
        "input": "",
        "output": "Tenses: present, past, future (simple, continuous, perfect, perfect continuous)",
        "text": "### Instruction:\nWhat is Tenses?\n\n### Response:\nTenses: present, past, future (simple, continuous, perfect, perfect continuous)"
    },
    {
        "instruction": "What is Article 12?",
        "input": "",
        "output": "Article 12: Definition of State",
        "text": "### Instruction:\nWhat is Article 12?\n\n### Response:\nArticle 12: Definition of State"
    },
    {
        "instruction": "What is the formula for Classification?",
        "input": "",
        "output": "- Recall = TP / (TP + FN) - out of actual positives, how many found",
        "text": "### Instruction:\nWhat is the formula for Classification?\n\n### Response:\n- Recall = TP / (TP + FN) - out of actual positives, how many found"
    },
    {
        "instruction": "What is 4. Semi-supervised?",
        "input": "",
        "output": "4. Semi-supervised: mix of labeled and unlabeled data",
        "text": "### Instruction:\nWhat is 4. Semi-supervised?\n\n### Response:\n4. Semi-supervised: mix of labeled and unlabeled data"
    },
    {
        "instruction": "What is SELECT?",
        "input": "",
        "output": "SELECT: retrieve data",
        "text": "### Instruction:\nWhat is SELECT?\n\n### Response:\nSELECT: retrieve data"
    },
    {
        "instruction": "What are the key concepts in Music Recommender?",
        "input": "",
        "output": "What: Personalized music recommendation system\nTech: Flask, Spotify API, scikit-learn\nKey Features:\n  - K-means clustering for music grouping\n  - Collaborative filtering for personalization\n  - Spotify API for music data\n  - 28% accuracy improvement over baseline\n  - Tested with 1000+ users\n  - Under 2 second inference\n\n================================================================================\nEND OF NOTES\n================================================================================",
        "text": "### Instruction:\nWhat are the key concepts in Music Recommender?\n\n### Response:\nWhat: Personalized music recommendation system\nTech: Flask, Spotify API, scikit-learn\nKey Features:\n  - K-means clustering for music grouping\n  - Collaborative filtering for personalization\n  - Spotify API for music data\n  - 28% accuracy improvement over baseline\n  - Tested with 1000+ users\n  - Under 2 second inference\n\n================================================================================\nEND OF NOTES\n================================================================================"
    },
    {
        "instruction": "What is Loss Asset?",
        "input": "",
        "output": "Loss Asset: uncollectible NPA",
        "text": "### Instruction:\nWhat is Loss Asset?\n\n### Response:\nLoss Asset: uncollectible NPA"
    },
    {
        "instruction": "What is NEFT?",
        "input": "",
        "output": "NEFT: National Electronic Funds Transfer (settled in batches)",
        "text": "### Instruction:\nWhat is NEFT?\n\n### Response:\nNEFT: National Electronic Funds Transfer (settled in batches)"
    },
    {
        "instruction": "What is ADB?",
        "input": "",
        "output": "ADB: Asian Development Bank - promotes development in Asia",
        "text": "### Instruction:\nWhat is ADB?\n\n### Response:\nADB: Asian Development Bank - promotes development in Asia"
    },
    {
        "instruction": "What is Used for?",
        "input": "",
        "output": "Used for: image classification, object detection, computer vision",
        "text": "### Instruction:\nWhat is Used for?\n\n### Response:\nUsed for: image classification, object detection, computer vision"
    },
    {
        "instruction": "What is SIDBI?",
        "input": "",
        "output": "SIDBI: Small Industries Development Bank of India",
        "text": "### Instruction:\nWhat is SIDBI?\n\n### Response:\nSIDBI: Small Industries Development Bank of India"
    },
    {
        "instruction": "What is Central Bank?",
        "input": "",
        "output": "Central Bank: RBI (regulates all banks)",
        "text": "### Instruction:\nWhat is Central Bank?\n\n### Response:\nCentral Bank: RBI (regulates all banks)"
    },
    {
        "instruction": "What is 2. Dockerfile uses dnf (RHEL) but base image is python?",
        "input": "",
        "output": "2. Dockerfile uses dnf (RHEL) but base image is python: 3.10-slim (Debian)",
        "text": "### Instruction:\nWhat is 2. Dockerfile uses dnf (RHEL) but base image is python?\n\n### Response:\n2. Dockerfile uses dnf (RHEL) but base image is python: 3.10-slim (Debian)"
    },
    {
        "instruction": "What is Supabase?",
        "input": "",
        "output": "Supabase: open source Firebase alternative",
        "text": "### Instruction:\nWhat is Supabase?\n\n### Response:\nSupabase: open source Firebase alternative"
    },
    {
        "instruction": "What is Machine Learning?",
        "input": "",
        "output": "Machine Learning: systems that learn from data without explicit programming",
        "text": "### Instruction:\nWhat is Machine Learning?\n\n### Response:\nMachine Learning: systems that learn from data without explicit programming"
    },
    {
        "instruction": "What is IMPS?",
        "input": "",
        "output": "IMPS: Immediate Payment Service (24x7, instant, up to 5 lakh)",
        "text": "### Instruction:\nWhat is IMPS?\n\n### Response:\nIMPS: Immediate Payment Service (24x7, instant, up to 5 lakh)"
    },
    {
        "instruction": "What is the significance/value of Article 13?",
        "input": "",
        "output": "Article 13: Laws inconsistent with Fundamental Rights are void",
        "text": "### Instruction:\nWhat is the significance/value of Article 13?\n\n### Response:\nArticle 13: Laws inconsistent with Fundamental Rights are void"
    },
    {
        "instruction": "What is the formula for P(A) + P(not A) = 1?",
        "input": "",
        "output": "For independent events: P(A and B) = P(A) x P(B)",
        "text": "### Instruction:\nWhat is the formula for P(A) + P(not A) = 1?\n\n### Response:\nFor independent events: P(A and B) = P(A) x P(B)"
    },
    {
        "instruction": "What is Information Gain?",
        "input": "",
        "output": "Information Gain: entropy-based measure",
        "text": "### Instruction:\nWhat is Information Gain?\n\n### Response:\nInformation Gain: entropy-based measure"
    },
    {
        "instruction": "What is SECTION 3?",
        "input": "",
        "output": "SECTION 3: ENGLISH LANGUAGE",
        "text": "### Instruction:\nWhat is SECTION 3?\n\n### Response:\nSECTION 3: ENGLISH LANGUAGE"
    },
    {
        "instruction": "What is the significance/value of Article 23?",
        "input": "",
        "output": "Article 23: Prohibition of traffic in human beings",
        "text": "### Instruction:\nWhat is the significance/value of Article 23?\n\n### Response:\nArticle 23: Prohibition of traffic in human beings"
    },
    {
        "instruction": "What is Shift coding?",
        "input": "",
        "output": "Shift coding: each letter shifted by fixed number",
        "text": "### Instruction:\nWhat is Shift coding?\n\n### Response:\nShift coding: each letter shifted by fixed number"
    },
    {
        "instruction": "What is the significance/value of Article 112?",
        "input": "",
        "output": "Article 112: Annual Financial Statement (Union Budget)",
        "text": "### Instruction:\nWhat is the significance/value of Article 112?\n\n### Response:\nArticle 112: Annual Financial Statement (Union Budget)"
    },
    {
        "instruction": "What is cv2.CascadeClassifier()?",
        "input": "",
        "output": "cv2.CascadeClassifier(): Haar cascade for face/eye detection",
        "text": "### Instruction:\nWhat is cv2.CascadeClassifier()?\n\n### Response:\ncv2.CascadeClassifier(): Haar cascade for face/eye detection"
    },
    {
        "instruction": "Explain your NeuroSpeech project.",
        "input": "",
        "output": "AI speech therapy platform using OpenAI Whisper for real-time transcription,\nIPA phonetic analysis for pronunciation feedback, Supabase for authentication\nand data storage with RLS for security, Realtime subscriptions for live\nleaderboard, Edge Functions for AI feedback generation, deployed on HuggingFace.",
        "text": "### Instruction:\nExplain your NeuroSpeech project.\n\n### Response:\nAI speech therapy platform using OpenAI Whisper for real-time transcription,\nIPA phonetic analysis for pronunciation feedback, Supabase for authentication\nand data storage with RLS for security, Realtime subscriptions for live\nleaderboard, Edge Functions for AI feedback generation, deployed on HuggingFace."
    },
    {
        "instruction": "What is Cross-validation?",
        "input": "",
        "output": "Cross-validation: k-fold, split data into k parts, train on k-1, test on 1",
        "text": "### Instruction:\nWhat is Cross-validation?\n\n### Response:\nCross-validation: k-fold, split data into k parts, train on k-1, test on 1"
    },
    {
        "instruction": "What is FTP?",
        "input": "",
        "output": "FTP: File Transfer Protocol",
        "text": "### Instruction:\nWhat is FTP?\n\n### Response:\nFTP: File Transfer Protocol"
    },
    {
        "instruction": "What is Softmax?",
        "input": "",
        "output": "Softmax: multi-class classification, outputs sum to 1",
        "text": "### Instruction:\nWhat is Softmax?\n\n### Response:\nSoftmax: multi-class classification, outputs sum to 1"
    },
    {
        "instruction": "What is Rational numbers?",
        "input": "",
        "output": "Rational numbers: can be expressed as p/q where q is not zero",
        "text": "### Instruction:\nWhat is Rational numbers?\n\n### Response:\nRational numbers: can be expressed as p/q where q is not zero"
    },
    {
        "instruction": "What is the formula for Duplicate ratio of a:b is a^2:b^2?",
        "input": "",
        "output": "Proportion: a:b = c:d means a, b, c, d are in proportion",
        "text": "### Instruction:\nWhat is the formula for Duplicate ratio of a:b is a^2:b^2?\n\n### Response:\nProportion: a:b = c:d means a, b, c, d are in proportion"
    },
    {
        "instruction": "What is Right triangle?",
        "input": "",
        "output": "Right triangle: Area = 1/2 x base x perpendicular",
        "text": "### Instruction:\nWhat is Right triangle?\n\n### Response:\nRight triangle: Area = 1/2 x base x perpendicular"
    },
    {
        "instruction": "What are the key concepts in Cybersecurity?",
        "input": "",
        "output": "Malware: malicious software\nVirus: replicates by infecting files\nTrojan: disguises as legitimate software\nRansomware: encrypts files and demands ransom\nPhishing: fake emails/websites to steal credentials\nEncryption: converting data to unreadable format\nTwo-factor authentication: extra security layer beyond password\n\n================================================================================\nSECTION 7: AI/ML INTERVIEW PREPARATION\n================================================================================",
        "text": "### Instruction:\nWhat are the key concepts in Cybersecurity?\n\n### Response:\nMalware: malicious software\nVirus: replicates by infecting files\nTrojan: disguises as legitimate software\nRansomware: encrypts files and demands ransom\nPhishing: fake emails/websites to steal credentials\nEncryption: converting data to unreadable format\nTwo-factor authentication: extra security layer beyond password\n\n================================================================================\nSECTION 7: AI/ML INTERVIEW PREPARATION\n================================================================================"
    },
    {
        "instruction": "What is EXIM Bank?",
        "input": "",
        "output": "EXIM Bank: Export-Import Bank of India",
        "text": "### Instruction:\nWhat is EXIM Bank?\n\n### Response:\nEXIM Bank: Export-Import Bank of India"
    },
    {
        "instruction": "What is Padding?",
        "input": "",
        "output": "Padding: adds zeros around input to preserve spatial dimensions",
        "text": "### Instruction:\nWhat is Padding?\n\n### Response:\nPadding: adds zeros around input to preserve spatial dimensions"
    },
    {
        "instruction": "What is Per capita income?",
        "input": "",
        "output": "Per capita income: NNP / Population",
        "text": "### Instruction:\nWhat is Per capita income?\n\n### Response:\nPer capita income: NNP / Population"
    },
    {
        "instruction": "What is AlexNet?",
        "input": "",
        "output": "AlexNet: won ImageNet 2012, made deep learning popular",
        "text": "### Instruction:\nWhat is AlexNet?\n\n### Response:\nAlexNet: won ImageNet 2012, made deep learning popular"
    },
    {
        "instruction": "What is the formula for Difference between CI and SI for 3 years = P x (R/100)^2 x (R/100 + 3)?",
        "input": "",
        "output": "Speed = Distance / Time",
        "text": "### Instruction:\nWhat is the formula for Difference between CI and SI for 3 years = P x (R/100)^2 x (R/100 + 3)?\n\n### Response:\nSpeed = Distance / Time"
    },
    {
        "instruction": "What are the key concepts in Deep Learning?",
        "input": "",
        "output": "Neural Network: layers of interconnected neurons\n  - Input layer, hidden layers, output layer\n  - Each connection has a weight\n\nActivation Functions:\n  - Sigmoid: 1/(1+e^-x), output 0 to 1, vanishing gradient problem\n  - ReLU: max(0,x), most popular, solves vanishing gradient\n  - Leaky ReLU: allows small negative values\n  - Tanh: output -1 to 1, zero-centered\n  - Softmax: multi-class classification, outputs sum to 1\n\nGradient Descent: optimization algorithm to minimize loss\n  - Batch Gradient Descent: uses all data, slow but stable\n  - Stochastic Gradient Descent (SGD): one sample at a time, n",
        "text": "### Instruction:\nWhat are the key concepts in Deep Learning?\n\n### Response:\nNeural Network: layers of interconnected neurons\n  - Input layer, hidden layers, output layer\n  - Each connection has a weight\n\nActivation Functions:\n  - Sigmoid: 1/(1+e^-x), output 0 to 1, vanishing gradient problem\n  - ReLU: max(0,x), most popular, solves vanishing gradient\n  - Leaky ReLU: allows small negative values\n  - Tanh: output -1 to 1, zero-centered\n  - Softmax: multi-class classification, outputs sum to 1\n\nGradient Descent: optimization algorithm to minimize loss\n  - Batch Gradient Descent: uses all data, slow but stable\n  - Stochastic Gradient Descent (SGD): one sample at a time, n"
    },
    {
        "instruction": "What is Ransomware?",
        "input": "",
        "output": "Ransomware: encrypts files and demands ransom",
        "text": "### Instruction:\nWhat is Ransomware?\n\n### Response:\nRansomware: encrypts files and demands ransom"
    },
    {
        "instruction": "What is NNP?",
        "input": "",
        "output": "NNP: Net National Product = GNP - Depreciation",
        "text": "### Instruction:\nWhat is NNP?\n\n### Response:\nNNP: Net National Product = GNP - Depreciation"
    },
    {
        "instruction": "What is Feed-forward network?",
        "input": "",
        "output": "Feed-forward network: applied to each position",
        "text": "### Instruction:\nWhat is Feed-forward network?\n\n### Response:\nFeed-forward network: applied to each position"
    },
    {
        "instruction": "What are the key concepts in MLOps and Deployment?",
        "input": "",
        "output": "Model Serialization: saving trained models\n  - TensorFlow: model.save(), tf.keras.models.load_model()\n  - PyTorch: torch.save(), torch.load()\n  - Scikit-learn: pickle, joblib\nFastAPI: fast web framework for building ML APIs (Parth uses this)\n  - @app.post(), @app.get(), Pydantic models for validation\n  - Used in Drive-Awake for drowsiness detection API\nDocker: containerize ML applications\n  - Dockerfile, docker build, docker run\n  - FROM, RUN, COPY, CMD instructions\nHuggingFace: platform for sharing ML models and spaces\n  - Parth's NeuroSpeech is deployed on HuggingFace Spaces\n  - transformers",
        "text": "### Instruction:\nWhat are the key concepts in MLOps and Deployment?\n\n### Response:\nModel Serialization: saving trained models\n  - TensorFlow: model.save(), tf.keras.models.load_model()\n  - PyTorch: torch.save(), torch.load()\n  - Scikit-learn: pickle, joblib\nFastAPI: fast web framework for building ML APIs (Parth uses this)\n  - @app.post(), @app.get(), Pydantic models for validation\n  - Used in Drive-Awake for drowsiness detection API\nDocker: containerize ML applications\n  - Dockerfile, docker build, docker run\n  - FROM, RUN, COPY, CMD instructions\nHuggingFace: platform for sharing ML models and spaces\n  - Parth's NeuroSpeech is deployed on HuggingFace Spaces\n  - transformers"
    },
    {
        "instruction": "What is 1 May?",
        "input": "",
        "output": "1 May: Labour Day / Maharashtra Day / Gujarat Day",
        "text": "### Instruction:\nWhat is 1 May?\n\n### Response:\n1 May: Labour Day / Maharashtra Day / Gujarat Day"
    },
    {
        "instruction": "What is values to 0-1, making gradients tiny. Solution?",
        "input": "",
        "output": "values to 0-1, making gradients tiny. Solution: ReLU activation, residual",
        "text": "### Instruction:\nWhat is values to 0-1, making gradients tiny. Solution?\n\n### Response:\nvalues to 0-1, making gradients tiny. Solution: ReLU activation, residual"
    },
    {
        "instruction": "What is the significance/value of Article 17?",
        "input": "",
        "output": "Article 17: Abolition of untouchability",
        "text": "### Instruction:\nWhat is the significance/value of Article 17?\n\n### Response:\nArticle 17: Abolition of untouchability"
    },
    {
        "instruction": "What is World Bank?",
        "input": "",
        "output": "World Bank: provides loans for development projects",
        "text": "### Instruction:\nWhat is World Bank?\n\n### Response:\nWorld Bank: provides loans for development projects"
    },
    {
        "instruction": "What is the formula for Speed = Distance / Time?",
        "input": "",
        "output": "Average Speed = Total Distance / Total Time",
        "text": "### Instruction:\nWhat is the formula for Speed = Distance / Time?\n\n### Response:\nAverage Speed = Total Distance / Total Time"
    },
    {
        "instruction": "What is the formula for Sum = Average x Number?",
        "input": "",
        "output": "Weighted average = (w1*x1 + w2*x2 + ...) / (w1 + w2 + ...)",
        "text": "### Instruction:\nWhat is the formula for Sum = Average x Number?\n\n### Response:\nWeighted average = (w1*x1 + w2*x2 + ...) / (w1 + w2 + ...)"
    },
    {
        "instruction": "What is Gradient Descent?",
        "input": "",
        "output": "Gradient Descent: optimization algorithm to minimize loss",
        "text": "### Instruction:\nWhat is Gradient Descent?\n\n### Response:\nGradient Descent: optimization algorithm to minimize loss"
    },
    {
        "instruction": "What is the formula for SECTION 4: BANKING AND FINANCIAL AWARENESS?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for SECTION 4: BANKING AND FINANCIAL AWARENESS?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "A train travels 240 km in 4 hours. What is its speed in km/h?",
        "output": "Speed = Distance / Time = 240 / 4 = 60 km/h",
        "text": "### Instruction:\nA train travels 240 km in 4 hours. What is its speed in km/h?\n\n### Response:\nSpeed = Distance / Time = 240 / 4 = 60 km/h"
    },
    {
        "instruction": "What is Batch Gradient Descent?",
        "input": "",
        "output": "Batch Gradient Descent: uses all data, slow but stable",
        "text": "### Instruction:\nWhat is Batch Gradient Descent?\n\n### Response:\nBatch Gradient Descent: uses all data, slow but stable"
    },
    {
        "instruction": "What is the formula for - Tested with 1000+ users?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for - Tested with 1000+ users?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is Triangle?",
        "input": "",
        "output": "Triangle: Area = 1/2 x base x height",
        "text": "### Instruction:\nWhat is Triangle?\n\n### Response:\nTriangle: Area = 1/2 x base x height"
    },
    {
        "instruction": "What is Emotion Detector?",
        "input": "",
        "output": "Emotion Detector: CNN for facial emotion recognition",
        "text": "### Instruction:\nWhat is Emotion Detector?\n\n### Response:\nEmotion Detector: CNN for facial emotion recognition"
    },
    {
        "instruction": "What is the formula for Sub-duplicate ratio of a:b is sqrt(a):sqrt(b)?",
        "input": "",
        "output": "Mean proportion of a and b = sqrt(ab)",
        "text": "### Instruction:\nWhat is the formula for Sub-duplicate ratio of a:b is sqrt(a):sqrt(b)?\n\n### Response:\nMean proportion of a and b = sqrt(ab)"
    },
    {
        "instruction": "If the simple interest on Rs 5000 for 2 years at 8% per annum is?",
        "output": "SI = (P x R x T) / 100 = (5000 x 8 x 2) / 100 = Rs 800",
        "text": "### Instruction:\nIf the simple interest on Rs 5000 for 2 years at 8% per annum is?\n\n### Response:\nSI = (P x R x T) / 100 = (5000 x 8 x 2) / 100 = Rs 800"
    },
    {
        "instruction": "What is GPT?",
        "input": "",
        "output": "GPT: Generative Pre-trained Transformer, autoregressive generation",
        "text": "### Instruction:\nWhat is GPT?\n\n### Response:\nGPT: Generative Pre-trained Transformer, autoregressive generation"
    },
    {
        "instruction": "What is Article 44?",
        "input": "",
        "output": "Article 44: Uniform Civil Code (Directive Principle)",
        "text": "### Instruction:\nWhat is Article 44?\n\n### Response:\nArticle 44: Uniform Civil Code (Directive Principle)"
    },
    {
        "instruction": "What is CPU?",
        "input": "",
        "output": "CPU: Central Processing Unit - brain of computer",
        "text": "### Instruction:\nWhat is CPU?\n\n### Response:\nCPU: Central Processing Unit - brain of computer"
    },
    {
        "instruction": "What is K-Means Clustering?",
        "input": "",
        "output": "K-Means Clustering: unsupervised, groups data into k clusters",
        "text": "### Instruction:\nWhat is K-Means Clustering?\n\n### Response:\nK-Means Clustering: unsupervised, groups data into k clusters"
    },
    {
        "instruction": "What are the key concepts in NLP Concepts?",
        "input": "",
        "output": "Tokenization: splitting text into tokens (words, subwords, characters)\nEmbedding: mapping tokens to dense vectors (Word2Vec, GloVe, FastText)\nTF-IDF: Term Frequency - Inverse Document Frequency\n  - TF: how often word appears in document\n  - IDF: how rare word is across all documents\nSentiment Analysis: classify text as positive/negative/neutral\nNamed Entity Recognition (NER): identify names, places, organizations\nText Classification: assign categories to text\n\nUsed in Parth's Projects:\n  - NeuroSpeech: OpenAI Whisper for speech-to-text (transformer-based)\n  - IPA phonetic analysis for pronunci",
        "text": "### Instruction:\nWhat are the key concepts in NLP Concepts?\n\n### Response:\nTokenization: splitting text into tokens (words, subwords, characters)\nEmbedding: mapping tokens to dense vectors (Word2Vec, GloVe, FastText)\nTF-IDF: Term Frequency - Inverse Document Frequency\n  - TF: how often word appears in document\n  - IDF: how rare word is across all documents\nSentiment Analysis: classify text as positive/negative/neutral\nNamed Entity Recognition (NER): identify names, places, organizations\nText Classification: assign categories to text\n\nUsed in Parth's Projects:\n  - NeuroSpeech: OpenAI Whisper for speech-to-text (transformer-based)\n  - IPA phonetic analysis for pronunci"
    },
    {
        "instruction": "What is LAN?",
        "input": "",
        "output": "LAN: Local Area Network - small area (office, building)",
        "text": "### Instruction:\nWhat is LAN?\n\n### Response:\nLAN: Local Area Network - small area (office, building)"
    },
    {
        "instruction": "What is the significance/value of Article 51A?",
        "input": "",
        "output": "Article 51A: Fundamental Duties (11 duties added by 42nd amendment)",
        "text": "### Instruction:\nWhat is the significance/value of Article 51A?\n\n### Response:\nArticle 51A: Fundamental Duties (11 duties added by 42nd amendment)"
    },
    {
        "instruction": "What is SVM?",
        "input": "",
        "output": "SVM: Support Vector Machine, finds optimal hyperplane between classes",
        "text": "### Instruction:\nWhat is SVM?\n\n### Response:\nSVM: Support Vector Machine, finds optimal hyperplane between classes"
    },
    {
        "instruction": "What is Recurring Deposit?",
        "input": "",
        "output": "Recurring Deposit: fixed amount deposited monthly for fixed period",
        "text": "### Instruction:\nWhat is Recurring Deposit?\n\n### Response:\nRecurring Deposit: fixed amount deposited monthly for fixed period"
    },
    {
        "instruction": "What is the formula for 1 km/h = 5/18 m/s?",
        "input": "",
        "output": "Relative speed (opposite direction) = S1 + S2",
        "text": "### Instruction:\nWhat is the formula for 1 km/h = 5/18 m/s?\n\n### Response:\nRelative speed (opposite direction) = S1 + S2"
    },
    {
        "instruction": "What is Direct speech?",
        "input": "",
        "output": "Direct speech: exact words in quotes. Indirect: reported speech",
        "text": "### Instruction:\nWhat is Direct speech?\n\n### Response:\nDirect speech: exact words in quotes. Indirect: reported speech"
    },
    {
        "instruction": "What is 1. Supervised Learning?",
        "input": "",
        "output": "1. Supervised Learning: labeled data, predict output (classification, regression)",
        "text": "### Instruction:\nWhat is 1. Supervised Learning?\n\n### Response:\n1. Supervised Learning: labeled data, predict output (classification, regression)"
    },
    {
        "instruction": "What is Pradhan Mantri Mudra Yojana?",
        "input": "",
        "output": "Pradhan Mantri Mudra Yojana: loans to small businesses",
        "text": "### Instruction:\nWhat is Pradhan Mantri Mudra Yojana?\n\n### Response:\nPradhan Mantri Mudra Yojana: loans to small businesses"
    },
    {
        "instruction": "What is Article 51A?",
        "input": "",
        "output": "Article 51A: Fundamental Duties (11 duties added by 42nd amendment)",
        "text": "### Instruction:\nWhat is Article 51A?\n\n### Response:\nArticle 51A: Fundamental Duties (11 duties added by 42nd amendment)"
    },
    {
        "instruction": "What is Some A are B?",
        "input": "",
        "output": "Some A are B: at least one A is B",
        "text": "### Instruction:\nWhat is Some A are B?\n\n### Response:\nSome A are B: at least one A is B"
    },
    {
        "instruction": "What is UPI?",
        "input": "",
        "output": "UPI: Unified Payments Interface (real time, mobile based)",
        "text": "### Instruction:\nWhat is UPI?\n\n### Response:\nUPI: Unified Payments Interface (real time, mobile based)"
    },
    {
        "instruction": "What is Data augmentation?",
        "input": "",
        "output": "Data augmentation: flip, rotate, zoom, shift, brightness change",
        "text": "### Instruction:\nWhat is Data augmentation?\n\n### Response:\nData augmentation: flip, rotate, zoom, shift, brightness change"
    },
    {
        "instruction": "What is Broadcasting?",
        "input": "",
        "output": "Broadcasting: operations on different shaped arrays",
        "text": "### Instruction:\nWhat is Broadcasting?\n\n### Response:\nBroadcasting: operations on different shaped arrays"
    },
    {
        "instruction": "What is the formula for 18 April: World Heritage Day?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for 18 April: World Heritage Day?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is Self-attention?",
        "input": "",
        "output": "Self-attention: each token attends to all other tokens",
        "text": "### Instruction:\nWhat is Self-attention?\n\n### Response:\nSelf-attention: each token attends to all other tokens"
    },
    {
        "instruction": "What is the formula for Q: What is attention mechanism in transformers??",
        "input": "",
        "output": "its representation. Attention(Q,K,V) = softmax(QK^T / sqrt(d_k)) V",
        "text": "### Instruction:\nWhat is the formula for Q: What is attention mechanism in transformers??\n\n### Response:\nits representation. Attention(Q,K,V) = softmax(QK^T / sqrt(d_k)) V"
    },
    {
        "instruction": "What is RAM?",
        "input": "",
        "output": "RAM: Random Access Memory - temporary storage, volatile",
        "text": "### Instruction:\nWhat is RAM?\n\n### Response:\nRAM: Random Access Memory - temporary storage, volatile"
    },
    {
        "instruction": "What is Bagging?",
        "input": "",
        "output": "Bagging: train each tree on random subset of data",
        "text": "### Instruction:\nWhat is Bagging?\n\n### Response:\nBagging: train each tree on random subset of data"
    },
    {
        "instruction": "What is For mutually exclusive events?",
        "input": "",
        "output": "For mutually exclusive events: P(A or B) = P(A) + P(B)",
        "text": "### Instruction:\nWhat is For mutually exclusive events?\n\n### Response:\nFor mutually exclusive events: P(A or B) = P(A) + P(B)"
    },
    {
        "instruction": "What are the key concepts in Indian Constitution (Key Articles)?",
        "input": "",
        "output": "Article 1: India shall be a Union of States\nArticle 12: Definition of State\nArticle 13: Laws inconsistent with Fundamental Rights are void\nArticle 14: Equality before law\nArticle 15: Prohibition of discrimination\nArticle 16: Equality of opportunity in public employment\nArticle 17: Abolition of untouchability\nArticle 19: Freedom of speech, assembly, movement, profession\nArticle 20: Protection against self-incrimination\nArticle 21: Protection of life and personal liberty\nArticle 21A: Right to education (6-14 years)\nArticle 22: Protection against arbitrary arrest\nArticle 23: Prohibition of traffi",
        "text": "### Instruction:\nWhat are the key concepts in Indian Constitution (Key Articles)?\n\n### Response:\nArticle 1: India shall be a Union of States\nArticle 12: Definition of State\nArticle 13: Laws inconsistent with Fundamental Rights are void\nArticle 14: Equality before law\nArticle 15: Prohibition of discrimination\nArticle 16: Equality of opportunity in public employment\nArticle 17: Abolition of untouchability\nArticle 19: Freedom of speech, assembly, movement, profession\nArticle 20: Protection against self-incrimination\nArticle 21: Protection of life and personal liberty\nArticle 21A: Right to education (6-14 years)\nArticle 22: Protection against arbitrary arrest\nArticle 23: Prohibition of traffi"
    },
    {
        "instruction": "What is WAN?",
        "input": "",
        "output": "WAN: Wide Area Network - country/world level (Internet is WAN)",
        "text": "### Instruction:\nWhat is WAN?\n\n### Response:\nWAN: Wide Area Network - country/world level (Internet is WAN)"
    },
    {
        "instruction": "What is KNN?",
        "input": "",
        "output": "KNN: K-Nearest Neighbors, classify based on k nearest data points",
        "text": "### Instruction:\nWhat is KNN?\n\n### Response:\nKNN: K-Nearest Neighbors, classify based on k nearest data points"
    },
    {
        "instruction": "What is the formula for - cv2.cvtColor(): convert color space (BGR to RGB, BGR to Gray)?",
        "input": "",
        "output": "- Eye Aspect Ratio (EAR) = (||p2-p6|| + ||p3-p5||) / (2 * ||p1-p4||)",
        "text": "### Instruction:\nWhat is the formula for - cv2.cvtColor(): convert color space (BGR to RGB, BGR to Gray)?\n\n### Response:\n- Eye Aspect Ratio (EAR) = (||p2-p6|| + ||p3-p5||) / (2 * ||p1-p4||)"
    },
    {
        "instruction": "What is Proportion?",
        "input": "",
        "output": "Proportion: a:b = c:d means a, b, c, d are in proportion",
        "text": "### Instruction:\nWhat is Proportion?\n\n### Response:\nProportion: a:b = c:d means a, b, c, d are in proportion"
    },
    {
        "instruction": "What is the significance/value of RBI nationalized?",
        "input": "",
        "output": "RBI nationalized: 1 January 1949",
        "text": "### Instruction:\nWhat is the significance/value of RBI nationalized?\n\n### Response:\nRBI nationalized: 1 January 1949"
    },
    {
        "instruction": "What is Overfitting?",
        "input": "",
        "output": "Overfitting: model performs well on training data but poorly on new data",
        "text": "### Instruction:\nWhat is Overfitting?\n\n### Response:\nOverfitting: model performs well on training data but poorly on new data"
    },
    {
        "instruction": "What is the formula for --- Inequalities ---?",
        "input": "",
        "output": "A >= B means A is greater than or equal to B",
        "text": "### Instruction:\nWhat is the formula for --- Inequalities ---?\n\n### Response:\nA >= B means A is greater than or equal to B"
    },
    {
        "instruction": "What is Bias-Variance Tradeoff?",
        "input": "",
        "output": "Bias-Variance Tradeoff: increasing complexity reduces bias but increases variance",
        "text": "### Instruction:\nWhat is Bias-Variance Tradeoff?\n\n### Response:\nBias-Variance Tradeoff: increasing complexity reduces bias but increases variance"
    },
    {
        "instruction": "What is Functions?",
        "input": "",
        "output": "Functions: monetary policy, currency issue, foreign exchange management",
        "text": "### Instruction:\nWhat is Functions?\n\n### Response:\nFunctions: monetary policy, currency issue, foreign exchange management"
    },
    {
        "instruction": "What is Encryption?",
        "input": "",
        "output": "Encryption: converting data to unreadable format",
        "text": "### Instruction:\nWhat is Encryption?\n\n### Response:\nEncryption: converting data to unreadable format"
    },
    {
        "instruction": "What are the key concepts in Time and Work?",
        "input": "",
        "output": "If A completes work in n days, A's 1 day work = 1/n\nIf A and B together complete work in n days, combined 1 day work = 1/n\nWork = Efficiency x Time\nIf A is twice as efficient as B, A takes half the time of B\nPipe and Cistern: inlet pipes fill, outlet pipes empty\nIf inlet fills in x hours and outlet empties in y hours:\n  Net filling rate = 1/x - 1/y (if x < y, tank fills)",
        "text": "### Instruction:\nWhat are the key concepts in Time and Work?\n\n### Response:\nIf A completes work in n days, A's 1 day work = 1/n\nIf A and B together complete work in n days, combined 1 day work = 1/n\nWork = Efficiency x Time\nIf A is twice as efficient as B, A takes half the time of B\nPipe and Cistern: inlet pipes fill, outlet pipes empty\nIf inlet fills in x hours and outlet empties in y hours:\n  Net filling rate = 1/x - 1/y (if x < y, tank fills)"
    },
    {
        "instruction": "What is NumPy?",
        "input": "",
        "output": "NumPy: numerical computing, arrays, matrix operations",
        "text": "### Instruction:\nWhat is NumPy?\n\n### Response:\nNumPy: numerical computing, arrays, matrix operations"
    },
    {
        "instruction": "What is the formula for If two trains of length L1 and L2 cross each other?",
        "input": "",
        "output": "1 km/h = 5/18 m/s",
        "text": "### Instruction:\nWhat is the formula for If two trains of length L1 and L2 cross each other?\n\n### Response:\n1 km/h = 5/18 m/s"
    },
    {
        "instruction": "What is Divisibility by 4?",
        "input": "",
        "output": "Divisibility by 4: last two digits divisible by 4",
        "text": "### Instruction:\nWhat is Divisibility by 4?\n\n### Response:\nDivisibility by 4: last two digits divisible by 4"
    },
    {
        "instruction": "What is NABARD?",
        "input": "",
        "output": "NABARD: National Bank for Agriculture and Rural Development",
        "text": "### Instruction:\nWhat is NABARD?\n\n### Response:\nNABARD: National Bank for Agriculture and Rural Development"
    },
    {
        "instruction": "What is the significance/value of Article 324?",
        "input": "",
        "output": "Article 324: Election Commission of India",
        "text": "### Instruction:\nWhat is the significance/value of Article 324?\n\n### Response:\nArticle 324: Election Commission of India"
    },
    {
        "instruction": "What is Sub-duplicate ratio of a?",
        "input": "",
        "output": "Sub-duplicate ratio of a: b is sqrt(a):sqrt(b)",
        "text": "### Instruction:\nWhat is Sub-duplicate ratio of a?\n\n### Response:\nSub-duplicate ratio of a: b is sqrt(a):sqrt(b)"
    },
    {
        "instruction": "What is Mini-batch?",
        "input": "",
        "output": "Mini-batch: uses small batches, balance between batch and SGD",
        "text": "### Instruction:\nWhat is Mini-batch?\n\n### Response:\nMini-batch: uses small batches, balance between batch and SGD"
    },
    {
        "instruction": "What is Fixed Deposit?",
        "input": "",
        "output": "Fixed Deposit: lump sum for fixed period at fixed interest",
        "text": "### Instruction:\nWhat is Fixed Deposit?\n\n### Response:\nFixed Deposit: lump sum for fixed period at fixed interest"
    },
    {
        "instruction": "What is 14 November?",
        "input": "",
        "output": "14 November: Children's Day (Nehru's birthday)",
        "text": "### Instruction:\nWhat is 14 November?\n\n### Response:\n14 November: Children's Day (Nehru's birthday)"
    },
    {
        "instruction": "What are the key concepts in Syllogism?",
        "input": "",
        "output": "All A are B: every A is included in B\nSome A are B: at least one A is B\nNo A is B: A and B have no overlap\nSome A are not B: at least one A is not B\nConclusion valid only if it follows necessarily from premises\nVenn diagrams help visualize syllogism problems",
        "text": "### Instruction:\nWhat are the key concepts in Syllogism?\n\n### Response:\nAll A are B: every A is included in B\nSome A are B: at least one A is B\nNo A is B: A and B have no overlap\nSome A are not B: at least one A is not B\nConclusion valid only if it follows necessarily from premises\nVenn diagrams help visualize syllogism problems"
    },
    {
        "instruction": "What is Convolutional Layer?",
        "input": "",
        "output": "Convolutional Layer: applies filters to detect features (edges, textures)",
        "text": "### Instruction:\nWhat is Convolutional Layer?\n\n### Response:\nConvolutional Layer: applies filters to detect features (edges, textures)"
    },
    {
        "instruction": "What is L2 (Ridge)?",
        "input": "",
        "output": "L2 (Ridge): adds squared weights to loss, prevents large weights",
        "text": "### Instruction:\nWhat is L2 (Ridge)?\n\n### Response:\nL2 (Ridge): adds squared weights to loss, prevents large weights"
    },
    {
        "instruction": "What is the formula for Check meaning: word must fit context?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for Check meaning: word must fit context?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is the significance/value of Article 356?",
        "input": "",
        "output": "Article 356: President's Rule in states",
        "text": "### Instruction:\nWhat is the significance/value of Article 356?\n\n### Response:\nArticle 356: President's Rule in states"
    },
    {
        "instruction": "What is the significance/value of Article 32?",
        "input": "",
        "output": "Article 32: Right to move Supreme Court (Dr. Ambedkar called it heart of Constitution)",
        "text": "### Instruction:\nWhat is the significance/value of Article 32?\n\n### Response:\nArticle 32: Right to move Supreme Court (Dr. Ambedkar called it heart of Constitution)"
    },
    {
        "instruction": "What is Shadow at sunset?",
        "input": "",
        "output": "Shadow at sunset: on East side",
        "text": "### Instruction:\nWhat is Shadow at sunset?\n\n### Response:\nShadow at sunset: on East side"
    },
    {
        "instruction": "What is IP Address?",
        "input": "",
        "output": "IP Address: unique address for every device on internet",
        "text": "### Instruction:\nWhat is IP Address?\n\n### Response:\nIP Address: unique address for every device on internet"
    },
    {
        "instruction": "What is Article 360?",
        "input": "",
        "output": "Article 360: Financial Emergency",
        "text": "### Instruction:\nWhat is Article 360?\n\n### Response:\nArticle 360: Financial Emergency"
    },
    {
        "instruction": "What is Development Banks?",
        "input": "",
        "output": "Development Banks: NABARD, NHB, SIDBI, EXIM Bank",
        "text": "### Instruction:\nWhat is Development Banks?\n\n### Response:\nDevelopment Banks: NABARD, NHB, SIDBI, EXIM Bank"
    },
    {
        "instruction": "What is FastAPI?",
        "input": "",
        "output": "FastAPI: fast web framework for building ML APIs (Parth uses this)",
        "text": "### Instruction:\nWhat is FastAPI?\n\n### Response:\nFastAPI: fast web framework for building ML APIs (Parth uses this)"
    },
    {
        "instruction": "What is Article 280?",
        "input": "",
        "output": "Article 280: Finance Commission",
        "text": "### Instruction:\nWhat is Article 280?\n\n### Response:\nArticle 280: Finance Commission"
    },
    {
        "instruction": "What is Mistral?",
        "input": "",
        "output": "Mistral: efficient transformer, sliding window attention, grouped query attention",
        "text": "### Instruction:\nWhat is Mistral?\n\n### Response:\nMistral: efficient transformer, sliding window attention, grouped query attention"
    },
    {
        "instruction": "What are the key concepts in Mensuration?",
        "input": "",
        "output": "Rectangle: Area = length x breadth, Perimeter = 2(l+b)\nSquare: Area = side^2, Perimeter = 4 x side\nTriangle: Area = 1/2 x base x height\nRight triangle: Area = 1/2 x base x perpendicular\nCircle: Area = pi x r^2, Circumference = 2 x pi x r\nCylinder: Volume = pi x r^2 x h, Curved Surface Area = 2 x pi x r x h\nCone: Volume = 1/3 x pi x r^2 x h, Slant height = sqrt(r^2 + h^2)\nSphere: Volume = 4/3 x pi x r^3, Surface Area = 4 x pi x r^2\nCuboid: Volume = l x b x h, Surface Area = 2(lb + bh + lh)\nCube: Volume = a^3, Surface Area = 6a^2",
        "text": "### Instruction:\nWhat are the key concepts in Mensuration?\n\n### Response:\nRectangle: Area = length x breadth, Perimeter = 2(l+b)\nSquare: Area = side^2, Perimeter = 4 x side\nTriangle: Area = 1/2 x base x height\nRight triangle: Area = 1/2 x base x perpendicular\nCircle: Area = pi x r^2, Circumference = 2 x pi x r\nCylinder: Volume = pi x r^2 x h, Curved Surface Area = 2 x pi x r x h\nCone: Volume = 1/3 x pi x r^2 x h, Slant height = sqrt(r^2 + h^2)\nSphere: Volume = 4/3 x pi x r^3, Surface Area = 4 x pi x r^2\nCuboid: Volume = l x b x h, Surface Area = 2(lb + bh + lh)\nCube: Volume = a^3, Surface Area = 6a^2"
    },
    {
        "instruction": "What is Divisibility by 2?",
        "input": "",
        "output": "Divisibility by 2: last digit is 0, 2, 4, 6, or 8",
        "text": "### Instruction:\nWhat is Divisibility by 2?\n\n### Response:\nDivisibility by 2: last digit is 0, 2, 4, 6, or 8"
    },
    {
        "instruction": "What is the significance/value of Article 21A?",
        "input": "",
        "output": "Article 21A: Right to education (6-14 years)",
        "text": "### Instruction:\nWhat is the significance/value of Article 21A?\n\n### Response:\nArticle 21A: Right to education (6-14 years)"
    },
    {
        "instruction": "What is the formula for Amount = Principal x (1 + Rate/100)^Time?",
        "input": "",
        "output": "When compounded quarterly: Rate = R/4, Time = 4T",
        "text": "### Instruction:\nWhat is the formula for Amount = Principal x (1 + Rate/100)^Time?\n\n### Response:\nWhen compounded quarterly: Rate = R/4, Time = 4T"
    },
    {
        "instruction": "What is the formula for Loss = Cost Price - Selling Price?",
        "input": "",
        "output": "Selling Price = Cost Price x (1 + Profit%/100)",
        "text": "### Instruction:\nWhat is the formula for Loss = Cost Price - Selling Price?\n\n### Response:\nSelling Price = Cost Price x (1 + Profit%/100)"
    },
    {
        "instruction": "What is Impact?",
        "input": "",
        "output": "Impact: Accessible speech therapy without expensive therapists",
        "text": "### Instruction:\nWhat is Impact?\n\n### Response:\nImpact: Accessible speech therapy without expensive therapists"
    },
    {
        "instruction": "What is Virus?",
        "input": "",
        "output": "Virus: replicates by infecting files",
        "text": "### Instruction:\nWhat is Virus?\n\n### Response:\nVirus: replicates by infecting files"
    },
    {
        "instruction": "What is Model Serialization?",
        "input": "",
        "output": "Model Serialization: saving trained models",
        "text": "### Instruction:\nWhat is Model Serialization?\n\n### Response:\nModel Serialization: saving trained models"
    },
    {
        "instruction": "What is 15 August?",
        "input": "",
        "output": "15 August: Independence Day",
        "text": "### Instruction:\nWhat is 15 August?\n\n### Response:\n15 August: Independence Day"
    },
    {
        "instruction": "What are the key concepts in Common Interview Questions?",
        "input": "",
        "output": "Q: What is the difference between supervised and unsupervised learning?\nA: Supervised learning uses labeled data to predict outputs (classification/regression).\nUnsupervised learning finds hidden patterns in unlabeled data (clustering/dimensionality reduction).\n\nQ: How do you handle imbalanced datasets?\nA: Oversampling minority class (SMOTE), undersampling majority class,\nclass weights in loss function, use F1/AUC instead of accuracy.\n\nQ: What is regularization and why is it used?\nA: Regularization adds penalty to loss function to prevent overfitting.\nL1 (Lasso) produces sparse models. L2 (Rid",
        "text": "### Instruction:\nWhat are the key concepts in Common Interview Questions?\n\n### Response:\nQ: What is the difference between supervised and unsupervised learning?\nA: Supervised learning uses labeled data to predict outputs (classification/regression).\nUnsupervised learning finds hidden patterns in unlabeled data (clustering/dimensionality reduction).\n\nQ: How do you handle imbalanced datasets?\nA: Oversampling minority class (SMOTE), undersampling majority class,\nclass weights in loss function, use F1/AUC instead of accuracy.\n\nQ: What is regularization and why is it used?\nA: Regularization adds penalty to loss function to prevent overfitting.\nL1 (Lasso) produces sparse models. L2 (Rid"
    },
    {
        "instruction": "What are the key concepts in Indian Geography?",
        "input": "",
        "output": "India: 7th largest country by area, 2nd most populous\nCapital: New Delhi\nArea: 3.287 million sq km\nCoastline: 7,516 km\nHighest peak: Kangchenjunga (Sikkim) - 3rd highest in world\nLongest river: Ganga (but Indus is longest if including Pakistan part)\nLargest state by area: Rajasthan\nSmallest state by area: Goa\nLargest state by population: Uttar Pradesh\nWestern Ghats: runs parallel to western coast, biodiversity hotspot\nEastern Ghats: discontinuous range along eastern coast\nDeccan Plateau: largest plateau in India\nThar Desert: in Rajasthan, only hot desert in India\nMajor rivers: Ganga, Yamuna, B",
        "text": "### Instruction:\nWhat are the key concepts in Indian Geography?\n\n### Response:\nIndia: 7th largest country by area, 2nd most populous\nCapital: New Delhi\nArea: 3.287 million sq km\nCoastline: 7,516 km\nHighest peak: Kangchenjunga (Sikkim) - 3rd highest in world\nLongest river: Ganga (but Indus is longest if including Pakistan part)\nLargest state by area: Rajasthan\nSmallest state by area: Goa\nLargest state by population: Uttar Pradesh\nWestern Ghats: runs parallel to western coast, biodiversity hotspot\nEastern Ghats: discontinuous range along eastern coast\nDeccan Plateau: largest plateau in India\nThar Desert: in Rajasthan, only hot desert in India\nMajor rivers: Ganga, Yamuna, B"
    },
    {
        "instruction": "What is the formula for SECTION 6: COMPUTER AWARENESS (FOR BANKING EXAMS)?",
        "input": "",
        "output": "================================================================================",
        "text": "### Instruction:\nWhat is the formula for SECTION 6: COMPUTER AWARENESS (FOR BANKING EXAMS)?\n\n### Response:\n================================================================================"
    },
    {
        "instruction": "What is Article 21?",
        "input": "",
        "output": "Article 21: Protection of life and personal liberty",
        "text": "### Instruction:\nWhat is Article 21?\n\n### Response:\nArticle 21: Protection of life and personal liberty"
    },
    {
        "instruction": "What is NACH?",
        "input": "",
        "output": "NACH: National Automated Clearing House (bulk transactions)",
        "text": "### Instruction:\nWhat is NACH?\n\n### Response:\nNACH: National Automated Clearing House (bulk transactions)"
    },
    {
        "instruction": "What is Article 32?",
        "input": "",
        "output": "Article 32: Right to move Supreme Court (Dr. Ambedkar called it heart of Constitution)",
        "text": "### Instruction:\nWhat is Article 32?\n\n### Response:\nArticle 32: Right to move Supreme Court (Dr. Ambedkar called it heart of Constitution)"
    },
    {
        "instruction": "What is Compound ratio of a?",
        "input": "",
        "output": "Compound ratio of a: b and c:d is ac:bd",
        "text": "### Instruction:\nWhat is Compound ratio of a?\n\n### Response:\nCompound ratio of a: b and c:d is ac:bd"
    },
    {
        "instruction": "What is Marginal Standing Facility (MSF)?",
        "input": "",
        "output": "Marginal Standing Facility (MSF): emergency borrowing from RBI above repo rate",
        "text": "### Instruction:\nWhat is Marginal Standing Facility (MSF)?\n\n### Response:\nMarginal Standing Facility (MSF): emergency borrowing from RBI above repo rate"
    },
    {
        "instruction": "What is the formula for If A is twice as efficient as B, A takes half the time of B?",
        "input": "",
        "output": "Net filling rate = 1/x - 1/y (if x < y, tank fills)",
        "text": "### Instruction:\nWhat is the formula for If A is twice as efficient as B, A takes half the time of B?\n\n### Response:\nNet filling rate = 1/x - 1/y (if x < y, tank fills)"
    },
    {
        "instruction": "What is Area?",
        "input": "",
        "output": "Area: 3.287 million sq km",
        "text": "### Instruction:\nWhat is Area?\n\n### Response:\nArea: 3.287 million sq km"
    },
    {
        "instruction": "What is RNN (Recurrent Neural Network)?",
        "input": "",
        "output": "RNN (Recurrent Neural Network): processes sequences, has memory",
        "text": "### Instruction:\nWhat is RNN (Recurrent Neural Network)?\n\n### Response:\nRNN (Recurrent Neural Network): processes sequences, has memory"
    },
    {
        "instruction": "What is Docker?",
        "input": "",
        "output": "Docker: containerize ML applications",
        "text": "### Instruction:\nWhat is Docker?\n\n### Response:\nDocker: containerize ML applications"
    },
    {
        "instruction": "What is Problem?",
        "input": "",
        "output": "Problem: vanishing gradient for long sequences",
        "text": "### Instruction:\nWhat is Problem?\n\n### Response:\nProblem: vanishing gradient for long sequences"
    },
    {
        "instruction": "What is DNS?",
        "input": "",
        "output": "DNS: Domain Name System - converts domain names to IP addresses",
        "text": "### Instruction:\nWhat is DNS?\n\n### Response:\nDNS: Domain Name System - converts domain names to IP addresses"
    },
    {
        "instruction": "What is Hard Disk?",
        "input": "",
        "output": "Hard Disk: permanent storage",
        "text": "### Instruction:\nWhat is Hard Disk?\n\n### Response:\nHard Disk: permanent storage"
    },
    {
        "instruction": "What is Same padding?",
        "input": "",
        "output": "Same padding: output same size as input",
        "text": "### Instruction:\nWhat is Same padding?\n\n### Response:\nSame padding: output same size as input"
    },
    {
        "instruction": "What is Article 19?",
        "input": "",
        "output": "Article 19: Freedom of speech, assembly, movement, profession",
        "text": "### Instruction:\nWhat is Article 19?\n\n### Response:\nArticle 19: Freedom of speech, assembly, movement, profession"
    },
    {
        "instruction": "What are the key concepts in Inequalities?",
        "input": "",
        "output": "A > B means A is definitely greater than B\nA >= B means A is greater than or equal to B\nA = B means equal\nComplementary pairs: >, < and >=, <=\nIf A > B > C, then A > C is definitely true\nIf A > B and B >= C, then A > C is definitely true\n\n================================================================================\nSECTION 3: ENGLISH LANGUAGE\n================================================================================",
        "text": "### Instruction:\nWhat are the key concepts in Inequalities?\n\n### Response:\nA > B means A is definitely greater than B\nA >= B means A is greater than or equal to B\nA = B means equal\nComplementary pairs: >, < and >=, <=\nIf A > B > C, then A > C is definitely true\nIf A > B and B >= C, then A > C is definitely true\n\n================================================================================\nSECTION 3: ENGLISH LANGUAGE\n================================================================================"
    },
    {
        "instruction": "What is Article 324?",
        "input": "",
        "output": "Article 324: Election Commission of India",
        "text": "### Instruction:\nWhat is Article 324?\n\n### Response:\nArticle 324: Election Commission of India"
    },
    {
        "instruction": "What is Statutory Liquidity Ratio (SLR)?",
        "input": "",
        "output": "Statutory Liquidity Ratio (SLR): percentage of deposits banks must maintain",
        "text": "### Instruction:\nWhat is Statutory Liquidity Ratio (SLR)?\n\n### Response:\nStatutory Liquidity Ratio (SLR): percentage of deposits banks must maintain"
    },
    {
        "instruction": "What is Steps for solving?",
        "input": "",
        "output": "Steps for solving: draw family tree, use M/F to denote gender",
        "text": "### Instruction:\nWhat is Steps for solving?\n\n### Response:\nSteps for solving: draw family tree, use M/F to denote gender"
    },
    {
        "instruction": "What is Rectangle?",
        "input": "",
        "output": "Rectangle: Area = length x breadth, Perimeter = 2(l+b)",
        "text": "### Instruction:\nWhat is Rectangle?\n\n### Response:\nRectangle: Area = length x breadth, Perimeter = 2(l+b)"
    },
    {
        "instruction": "What is Learning Rate?",
        "input": "",
        "output": "Learning Rate: step size for gradient descent",
        "text": "### Instruction:\nWhat is Learning Rate?\n\n### Response:\nLearning Rate: step size for gradient descent"
    },
    {
        "instruction": "What is Some A are not B?",
        "input": "",
        "output": "Some A are not B: at least one A is not B",
        "text": "### Instruction:\nWhat is Some A are not B?\n\n### Response:\nSome A are not B: at least one A is not B"
    },
    {
        "instruction": "What is 1 April?",
        "input": "",
        "output": "1 April: RBI Foundation Day",
        "text": "### Instruction:\nWhat is 1 April?\n\n### Response:\n1 April: RBI Foundation Day"
    },
    {
        "instruction": "What is Article 21A?",
        "input": "",
        "output": "Article 21A: Right to education (6-14 years)",
        "text": "### Instruction:\nWhat is Article 21A?\n\n### Response:\nArticle 21A: Right to education (6-14 years)"
    },
    {
        "instruction": "What is MAN?",
        "input": "",
        "output": "MAN: Metropolitan Area Network - city level",
        "text": "### Instruction:\nWhat is MAN?\n\n### Response:\nMAN: Metropolitan Area Network - city level"
    }
]