quiz.json

{
  "lesson": "28-alignment-research-ecosystem",
  "title": "Alignment Research Ecosystem - MATS, Redwood, Apollo, METR",
  "questions": [
    {
      "stage": "pre",
      "question": "What does MATS do, and what is its rough scale?",
      "options": [
        "A reward-model training service",
        "ML Alignment & Theory Scholars: a research mentorship program with 527+ scholars since 2021, 180+ papers, and roughly 80% of pre-2025 alumni working on safety/security",
        "A regulatory body that issues AI export licenses",
        "A frontier safety institute owned by the EU"
      ],
      "correct": 1,
      "explanation": ""
    },
    {
      "stage": "check",
      "question": "Which agenda did Redwood Research introduce?",
      "options": [
        "Differential privacy for LLMs",
        "AI Control (Lesson 10): safety despite subversion via U / T / H protocols",
        "Watermarking via SynthID",
        "Constitutional AI"
      ],
      "correct": 1,
      "explanation": ""
    },
    {
      "stage": "check",
      "question": "Which organisation authored the In-Context Scheming paper and the 'Towards Safety Cases for AI Scheming' framework?",
      "options": [
        "METR",
        "MATS",
        "Eleos AI",
        "Apollo Research"
      ],
      "correct": 3,
      "explanation": ""
    },
    {
      "stage": "check",
      "question": "What is METR's distinctive methodological style in the alignment ecosystem?",
      "options": [
        "Adversarial training of base models",
        "Task-based capability evaluations, autonomous-task time-horizon studies, and framework synthesis (e.g., 'Common Elements of Frontier AI Safety Policies')",
        "Hardware-attested verifiable inference",
        "Closed-source red-teaming only"
      ],
      "correct": 1,
      "explanation": ""
    },
    {
      "stage": "post",
      "question": "Why does multi-organisation co-authorship matter for safety claims?",
      "options": [
        "It increases citation counts",
        "It speeds up publication",
        "It is mandated by the EU AI Act",
        "Labs evaluating their own models have a structural conflict of interest; external evaluators (Redwood, Apollo, METR, Eleos, UK AISI) can raise and validate failure modes the lab might underreport"
      ],
      "correct": 3,
      "explanation": ""
    },
    {
      "stage": "post",
      "question": "What is Eleos AI Research's primary contribution to the ecosystem?",
      "options": [
        "AI Control benchmarks",
        "Watermarking standards",
        "California training-data law drafting",
        "Pre-deployment model-welfare evaluations (Lesson 19), including the external welfare assessment in Claude Opus 4's system card"
      ],
      "correct": 3,
      "explanation": ""
    }
  ]
}