{
  "holdout_size": 24,
  "correct": 23,
  "accuracy": 0.9583333333333334,
  "macro_f1": 0.9772727272727273,
  "confusion": {
    "lesson": {
      "lesson": 6,
      "roleplay": 0,
      "feedback": 0,
      "practice": 0,
      "clarify": 0
    },
    "roleplay": {
      "lesson": 0,
      "roleplay": 6,
      "feedback": 0,
      "practice": 0,
      "clarify": 0
    },
    "feedback": {
      "lesson": 0,
      "roleplay": 0,
      "feedback": 6,
      "practice": 0,
      "clarify": 0
    },
    "practice": {
      "lesson": 0,
      "roleplay": 0,
      "feedback": 0,
      "practice": 5,
      "clarify": 1
    }
  },
  "outcomes": [
    {
      "text": "I need a lesson outline for a B2 business class",
      "expected": "lesson",
      "prediction": "lesson"
    },
    {
      "text": "Can you plan an English lesson with a warm-up?",
      "expected": "lesson",
      "prediction": "lesson"
    },
    {
      "text": "Create a 45-minute classroom plan with a learning objective",
      "expected": "lesson",
      "prediction": "lesson"
    },
    {
      "text": "Monte uma aula A2 com uma atividade de revisão",
      "expected": "lesson",
      "prediction": "lesson"
    },
    {
      "text": "Planeje uma sequência didática para professores",
      "expected": "lesson",
      "prediction": "lesson"
    },
    {
      "text": "Preciso de um plano de aula sobre reservas",
      "expected": "lesson",
      "prediction": "lesson"
    },
    {
      "text": "Can we simulate a conversation at immigration?",
      "expected": "roleplay",
      "prediction": "roleplay"
    },
    {
      "text": "Play the customer and talk to me in English",
      "expected": "roleplay",
      "prediction": "roleplay"
    },
    {
      "text": "Start a hotel roleplay and wait for my reply",
      "expected": "roleplay",
      "prediction": "roleplay"
    },
    {
      "text": "Faça de conta que é o atendente e converse comigo",
      "expected": "roleplay",
      "prediction": "roleplay"
    },
    {
      "text": "Podemos encenar um diálogo no check-in?",
      "expected": "roleplay",
      "prediction": "roleplay"
    },
    {
      "text": "Vamos simular uma conversa para fazer um pedido",
      "expected": "roleplay",
      "prediction": "roleplay"
    },
    {
      "text": "Please review my sentence and tell me what to fix",
      "expected": "feedback",
      "prediction": "feedback"
    },
    {
      "text": "Explain the grammar error in what I wrote",
      "expected": "feedback",
      "prediction": "feedback"
    },
    {
      "text": "Is my English answer natural? Give feedback",
      "expected": "feedback",
      "prediction": "feedback"
    },
    {
      "text": "Pode corrigir minha frase e explicar a mudança?",
      "expected": "feedback",
      "prediction": "feedback"
    },
    {
      "text": "Quero saber onde errei na frase",
      "expected": "feedback",
      "prediction": "feedback"
    },
    {
      "text": "Revise meu inglês e mostre uma versão melhor",
      "expected": "feedback",
      "prediction": "feedback"
    },
    {
      "text": "Create a quiz with three options per question",
      "expected": "practice",
      "prediction": "practice"
    },
    {
      "text": "Give me a gap-fill exercise with an answer key",
      "expected": "practice",
      "prediction": "practice"
    },
    {
      "text": "I need some drills to practise vocabulary",
      "expected": "practice",
      "prediction": "clarify"
    },
    {
      "text": "Faça um exercício de lacunas com gabarito",
      "expected": "practice",
      "prediction": "practice"
    },
    {
      "text": "Quero treinar palavras com um quiz rápido",
      "expected": "practice",
      "prediction": "practice"
    },
    {
      "text": "Dê três questões de múltipla escolha",
      "expected": "practice",
      "prediction": "practice"
    }
  ],
  "train_size": 48
}
