Evaluator last updated June 15, 2026.
At a glance
| |
|---|
| Input type | Informational text |
| Passage length | 200 words or more |
| Supported grades | 3–12 |
Intertextuality is distinct from Subject Matter Knowledge:
- Subject Matter Knowledge – Content that a text assumes (facts, concepts, domain knowledge)
- Intertextuality - Relational knowledge that a text assumes (prior encounters with texts, references and conventions)
Model and prompt
For instructions on running the evaluator, see Quickstart.
| |
|---|
| Model used | gemini-3-flash-preview |
| Temperature | 0 |
| Prompts | View prompts ↗ |
| Python notebook | View notebook ↗ |
| Requirement | Supported | Required |
|---|
| Text passage | Informational text | Yes |
| Target grade level | Enables grade-specific guidance | No |
Output
| Field | Description |
|---|
complexity_score | Intertextual complexity level: slightly_complex, moderately_complex, very_complex, or exceedingly_complex. |
reasoning | Synthesized decision that explains the rating and relevant context, cites specific references and their impact on comprehension, etc. |
detailed_summary | Individual factors that drive the rating (specific references, allusions, genre conventions, and shared discourse detected, with descriptions and their effect on intertextual demand) |
assumptions_and_scaffolding | Analysis of what the author assumes the reader already knows versus what the text explains, with suggestions for making the demands accessible |
recommended_use_cases | Instructional opportunities for using the text and its intertextual demands strategically. |
Interpreting results
The evaluator returns one of the following ratings, along with reasoning, to help developers determine the best course of action.
Ratings are based on how many references or allusions are present, how central they are to meaning, and how widely accessible they are likely to be for students at the target grade.
| Rating | Meaning |
|---|
| Slightly complex | Few or no references; any allusions are explained in the text or widely familiar. Meaning does not depend on outside textual or cultural encounters. |
| Moderately complex | Some references, allusions, or genre conventions are present and accessible to grade-level readers. They enrich meaning but rarely block it. |
| Very complex | Multiple references are central to meaning; the text assumes familiarity with other texts, cultural touchstones, or genre conventions it does not introduce. |
| Exceedingly complex | Dense, layered references are essential to meaning; the text assumes deep familiarity with a body of texts, traditions, or discourse, and is largely inaccessible without prior encounters. |
See Grade Level
Appropriateness
for an evaluator that synthesizes all text complexity dimensions and makes an
overall grade-level judgment. Rubric
| Slightly complex | Moderately complex | Very complex | Exceedingly complex |
|---|
| Few or no intertextual references; any present are explained or widely familiar; meaning is self-contained. | Some references, allusions, or genre conventions; generally accessible at grade level; meaning is enriched but not dependent on them. | Multiple references central to meaning; assumes familiarity with texts, cultural touchstones, or genre conventions not introduced in the text. | Dense, layered references essential to meaning; assumes deep familiarity with a body of texts, traditions, or discourse not part of the text. |
Accuracy and validation
This evaluator is provided as Early access. Comprehensive accuracy
measures are still evolving, and validation testing is ongoing.
| Metric | Result |
|---|
| Complexity score accuracy | 85% agreement with expert annotations |
| Expert agreement | 80% (8 of 10 sampled outputs passed expert review; threshold 60%) |
| Reasoning soundness | Average 3.5 / 5 |
| Dataset source | CLEAR Corpus↗ |
The benchmark dataset does not yet include validated examples of “Very
complex” or “Exceedingly complex” texts, which are uncommon and difficult to
annotate consistently for grades 3–12. Use caution when applying this
evaluator to higher grade levels where such texts are more frequent.The evaluator is also tuned for informational texts and may perform less reliably
on literary texts.
Evaluator release history
| Date | Changes |
|---|
| June 16, 2026 | First release. |
