Glossary

Key terms and definitions for the InceptBench framework

Glossary

A comprehensive guide to key terms used throughout the InceptBench framework and educational content evaluation.

A

Article

A complete educational document combining multiple learning components (text, images, embedded questions) into a unified pedagogical experience. Formatted in markdown with hierarchical headings, sequential flow, and contextual integration of mixed-media elements.

B

Benchmark Methodology

A systematic approach using specialized evaluators to assess educational content generation systems. Unlike benchmark results, a methodology defines the framework and evaluation criteria used for assessment.

D

Direct Instruction (DI)

A highly structured, teacher-led instructional approach emphasizing clear, explicit teaching of specific skills in a systematic manner. One of the core pedagogical principles of Incept. Learn more →

E

EduBench (External)

An external open-source benchmark included in InceptBench for diversity and external baseline comparison. Evaluator name: external_edubench. Evaluates across 6 educational tasks (QA, EC, IP, AG, QG, TMG) with scores on a 0-10 scale, averaged then normalized to 0-1. Learn more →

Evaluator

The unified InceptBench evaluation system that automatically assesses educational content quality. The evaluator intelligently routes to specialized internal methods based on content type and parameters—no manual configuration required.

Evaluator Version

Version identifier for evaluators (e.g., v1.0.0) indicating the specific release and capabilities of an evaluation tool.

F

Final Score

An aggregated quality score (0-1 scale) calculated across all evaluators run on a specific question, providing an overall assessment of content quality.

G

Generated Article Schema

Input schema for article evaluation. Structure:

{
  "id": "<question_id>",
  "request": {
    "grade": "<grade>",
    "subject": "<subject>",
    "type": "article",
    "difficulty": "<difficulty>",
    "locale": "<locale>",
    "skills": {
      "lesson_title": "<lesson_title>",
      "substandard_id": "<substandard_id>",
      "substandard_description": "<substandard_description>"
    },
    "instruction": "<content_instruction>"
  },
  "content": {
    "content": "<content_in_markdown_format>",
    "additional_details": "<optional_additional_details>"
  }
}

Generated Fill In Schema

Input schema for fill-in question evaluation. Structure:

{
  "id": "<question_id>",
  "request": {
    "grade": "<grade>",
    "subject": "<subject>",
    "type": "fill-in",
    "difficulty": "<difficulty>",
    "locale": "<locale>",
    "skills": {
      "lesson_title": "<lesson_title>",
      "substandard_id": "<substandard_id>",
      "substandard_description": "<substandard_description>"
    },
    "instruction": "<content_instruction>"
  },
  "content": {
    "question": "<question>",
    "answer": "<correct_answer>",
    "answer_explanation": "<answer_explanation>",
    "image_url": ["<image_url_1>", "<image_url_2>", ...],
    "additional_details": "<optional_additional_details>"
  }
}

Generated Multiple Choice Question Schema

Input schema for question evaluation (MCQ). Structure:

{
  "id": "<question_id>",
  "request": {
    "grade": "<grade>",
    "subject": "<subject>",
    "type": "mcq",
    "difficulty": "<difficulty>",
    "locale": "<locale>",
    "skills": {
      "lesson_title": "<lesson_title>",
      "substandard_id": "<substandard_id>",
      "substandard_description": "<substandard_description>"
    },
    "instruction": "<content_instruction>"
  },
  "content": {
    "question": "<question>",
    "answer": "<correct_answer_key>",
    "answer_explanation": "<answer_explanation>",
    "answer_options": [
      { "key": "A", "text": "<option_1>" },
      { "key": "B", "text": "<option_2>" },
      { "key": "C", "text": "<option_3>" },
      { "key": "D", "text": "<option_4>" },
      { "key": "E", "text": "<option_5>" }
    ],
    "image_url": ["<image_url_1>", "<image_url_2>", ...],
    "additional_details": "<optional_additional_details>"
  }
}

I

Image Quality Evaluation

Automatic detection and assessment of educational images using Direct Instruction rubric-based scoring (v1.3.0). When any content includes an image_url, image quality evaluation is automatically enabled to ensure visual content meets pedagogical standards. Evaluates both images that accompany text (accompaniment mode) and standalone educational images.

InceptBench

A unified evaluation framework for educational content that automatically routes to specialized assessment methods based on content characteristics. One intelligent evaluator that handles all K-12 subjects and content types without manual configuration. Designed to be target-system agnostic. Current version: v1.4.0.

K

K-12

Kindergarten through 12th grade, representing the full span of primary and secondary education in many educational systems.

M

MCQ (Multiple Choice Question)

A question format presenting one correct answer alongside multiple distractors, requiring students to identify the correct option.

MTSS (Multi-Tiered System of Supports)

A comprehensive framework of evidence-based practices designed to meet the diverse academic and behavioral needs of all students through tiered interventions. Learn more →

P

Pedagogy

The method and practice of teaching, including instructional strategies, learning theories, and educational principles. Incept pedagogy is grounded in 8 core pillars. Learn more →

Pedagogical Value

A dimension assessing how well educational content promotes effective learning, critical thinking, and skill development.

Q

Quality Control (QC)

Systematic evaluation processes ensuring educational content meets defined standards for accuracy, clarity, and pedagogical effectiveness.

R

Recommendation

An evaluator output classifying content as “accept” (ready to use), “revise” (needs improvements), or “reject” (does not meet standards).

Request Schema

InceptBench API request structure supports two formats:

Simple Request (Plain Text):

{
  "content": "Your educational content as plain text",
  "curriculum": "common_core",
  "generation_prompt": "Optional prompt used to generate this content"
}

Structured Request (Detailed):

{
  "generated_content": [
    {
      "id": "<content_id>",
      "request": {
        "grade": "<grade>",
        "subject": "<subject>",
        "type": "<content_type>",
        "difficulty": "<difficulty>",
        "locale": "<locale>",
        "skills": {
          "lesson_title": "<lesson_title>",
          "substandard_id": "<substandard_id>",
          "substandard_description": "<substandard_description>"
        },
        "instruction": "<content_instruction>"
      },
      "content": {
        // Content structure varies by type (mcq, fill-in, article)
      }
    }
  ]
}

Parameters: grade (K, 1-12), subject (math, ela, science, social-studies, general), type (mcq, multi-correct, fill-in, match, article), difficulty (easy, medium, hard), locale (e.g., en-AE, ar-AE, en-IN). The evaluator automatically determines which internal methods to use based on your content and parameters. See Generated Question Schema, Generated Fill In Schema, and Generated Article Schema.

Routing Parameters

Optional parameters (subject, grade, type) that help InceptBench automatically select the most appropriate evaluation methods for your content. These parameters enable intelligent routing without manual configuration of evaluation methods.

Response Schema

InceptBench API response structure:

{
  "request_id": "uuid",
  "evaluations": {
    "q1": {
      "inceptbench_new_evaluation": {
        "content_type": "question",
        "overall": {"score": 0.85, "reasoning": "...", "suggested_improvements": "..."},
        "factual_accuracy": {"score": 1.0, "reasoning": "...", "suggested_improvements": null},
        "educational_accuracy": {"score": 1.0, "reasoning": "...", "suggested_improvements": null},
        "curriculum_alignment": {"score": 0.9, "reasoning": "...", "suggested_improvements": "..."},
        "difficulty_alignment": {"score": 0.0, "reasoning": "...", "suggested_improvements": "..."},
        "weighted_score": 0.8387
      },
      "score": 0.8
    },
    "text1": {
      "inceptbench_new_evaluation": {
        "content_type": "article",
        "overall": {"score": 0.62, "reasoning": "...", "suggested_improvements": "..."},
        "factual_accuracy": {"score": 1.0, "reasoning": "...", "suggested_improvements": null},
        "teaching_quality": {"score": 0.0, "reasoning": "...", "suggested_improvements": "..."},
        "worked_examples": {"score": 0.0, "reasoning": "...", "suggested_improvements": "..."},
        "weighted_score": 0.5
      },
      "score": 0.62
    },
    "passage1": {
      "inceptbench_new_evaluation": {
        "content_type": "nonfiction_reading",
        "overall": {"score": 0.64, "reasoning": "...", "suggested_improvements": "..."},
        "reading_level_match": {"score": 1.0, "reasoning": "...", "suggested_improvements": null},
        "weighted_score": 0.6522
      },
      "score": 0.64
    }
  },
  "evaluation_time_seconds": 119.67,
  "inceptbench_version": "x.y.z"
}

S

Scarborough’s Reading Rope

A research-based framework illustrating how multiple strands of language and literacy skills interweave to create skilled reading comprehension. Learn more →

Skill

A specific, well-defined learning objective or competency within a subject and grade level.

Skill Schema

Metadata schema for educational content. Structure:

{
  "lesson_title": "<lesson_title>",
  "substandard_id": "<substandard_id>",
  "substandard_description": "<substandard_description>"
}

Required fields: title, grade, subject. Defaults: subject="mathematics", difficulty="medium", language="en". Optional: description.

Subject

An academic discipline or area of study (e.g., Mathematics, Reading Comprehension, Science).

T

Target System

The educational content generation system being evaluated by InceptBench. The framework is designed to be system-agnostic and work with any K-12 content generator.

Text Content

Educational passages, explanations, and text materials that can be evaluated for pedagogical quality using InceptBench. Includes passages, explanations, and general educational text (distinct from questions).