Knowledge gap finder: how to identify misunderstandings and suggest exercises
The 'Knowledge gap finder' tool helps teachers identify misunderstandings in student answers and suggest targeted practice. In this article we will explore how to work with the two text fields, formulate source material and goals, and then verify and refine the obtained result. We will look at a concrete example with percentages, discuss typical errors and technical limitations. The article contains no guarantees and does not replace pedagogical expertise.
Open tool: Knowledge gap finder ↗How the tool works and what it can do
"Knowledge gap finder" is a web form with two text fields: 'Source material' and 'Goal or desired result'. In the first field you insert the student's answers, and in the second you describe what you want to get: for example, 'identify misunderstandings in the topic of percentages' or 'suggest exercises on discount addition'. The tool analyzes the text and returns a single result in a separate panel. You can copy it or download it in TXT format.
The tool works only with text. It does not support document uploads, OCR, audio or images. There is no batch processing, automatic saving or integrations. Each request must be entered anew. The fields have limits: 8000 characters for source material and 1000 characters for the goal. The result may contain errors, so it must be checked manually. Queue and availability may change.
It is important to understand that the tool has no difficulty, style or level of change settings. All nuances are set by text. The more precisely the goal is formulated, the more useful the result. Do not expect the tool to guess the context you have not described.
Source material: "I think percentages are added directly: a 20% discount, then another 10% — that is 30%.". Goal: "Identify misunderstandings and suggest exercises".
Formulating source material and goal
To get a substantive analysis, in the 'Source material' field it is better to insert not only the answer itself but also a brief context: class, topic, prior knowledge. For example: "7th grade student, topic percentages, previously studied fractions. Answer: ...". This will help the tool more accurately determine which specific concepts cause difficulties.
In the 'Goal or desired result' field try to avoid overly general formulations like 'help with mathematics'. The more specific the goal, the more useful the recommendations. For example: "Determine which concepts the student confuses, and suggest three exercises on sequential discounts." You can also indicate which errors you have already noticed and ask the tool to confirm or refute your hypothesis.
Remember that the tool interprets the goal as an instruction. If you want to get different analysis variants, change the goal formulation. For example, the goal 'find all errors' will give one list, while the goal 'explain why the student thinks that way' will give another. Narrow requests usually give more specific recommendations, but may miss related gaps.
Variant 1 (broad goal): "Identify all possible misunderstandings in the answer". Variant 2 (narrow goal): "Focus on the error of adding percentages and suggest one exercise".
How to get different analysis variants
The first way is to change the focus in the goal. For one and the same student answer you can request: "Determine which concepts the student confuses", "Suggest three exercises on sequential discounts" or "Explain how the student might reason correctly". Each request will give a different result, because the tool follows your instruction.
The second way is to add additional information to the source material. If you include not only the answer but also context, the analysis will become more accurate. You can also indicate which specific errors you have already noticed and ask the tool to confirm or refute your hypothesis.
The third way is to vary the goal formulation from broad to narrow. Broad: "Help understand what the student has not learned". Narrow: "Find a specific error in calculating the final price after two discounts". Narrow requests usually give more specific recommendations, but may miss related gaps.
Variant A: "The student does not understand that percentages are taken from different bases". Variant B: "Error in that the student adds percentages, not multiplies coefficients". Variant C: "The student confuses percentage of a number and percentage change".
Comparison and selection of the best variant
After you have received several results with different goals, compare them. Pay attention to which gaps are mentioned in each variant. If one and the same error appears in all variants (for example, adding percentages), this is a strong signal. If the variants contradict each other, the source material may have been insufficiently complete.
Also compare the practical assignments proposed. In one variant there may be exercises on calculation, in another — on explaining in one's own words, in a third — on finding errors in ready solutions. Diversity is useful: you can choose the tasks that best suit your student.
Choose the variant that most closely matches your pedagogical goal. If you want to quickly check understanding, a variant with specific exercises will suit. If you are preparing a detailed explanation, choose a variant with analysis of the student's reasoning. After choosing, refine the result manually: copy the text to your editor, add your own examples, adapt formulations to the student's age.
You chose variant B, but it contains too complex terms. Rewrite: "You add percentages, but each discount is calculated from the new price. Therefore 20% and 10% are not 30%, but 28% of the initial price". Add a numerical example.
Checking the result and typical errors
Check the result against the student's original answer. The tool may miss context or invent a gap that does not exist. Verify every statement against the text you entered. If the tool writes "the student does not know the multiplication table", but the answer is only about percentages, that is a false conclusion.
Also check that the recommendations do not contradict each other. For example, one exercise may require adding percentages, another — multiplying coefficients. Make sure the tasks are consistent and will not confuse the student.
A typical error is too general a goal. If you write "help with mathematics", the result will be vague. Another error is entering extraneous data in the source material that is unrelated to the task. This can confuse the tool. Try to give only the information that helps understand the student's answer.
Check: the tool suggested an exercise "Add 20% and 10%". This contradicts the correct approach. You must reject such an exercise and reformulate the goal.
Limitations and ethical considerations
The tool is not a replacement for a teacher. It only helps notice possible gaps. The model can make mistakes, so the final decision is made by the teacher. Do not completely trust the automatic analysis, especially in sensitive topics.
Technical limitations: no batch processing, automatic saving, integrations. You cannot upload a document or audio. Everything you enter is processed as text. The result can be downloaded only in TXT. Images are vector compositions (SVG, PNG), not photorealistic generation. Speech is eSpeak WAV, subtitles are SRT. No DOCX, PDF, OCR or voice cloning.
Ethical aspects: do not enter personal student data without consent. Remember that the tool does not give guarantees and does not bear responsibility. Use it as an auxiliary means for reflection. If necessary, anonymize the text before entering.
If you are working with a student, obtain consent to process their answers. Anonymize the text before entering.
The 'Knowledge gap finder' tool allows obtaining different analysis variants if you change the criteria and goals. Comparing several variants helps see stable gaps and choose the best. However the result requires verification and refinement. Take into account limitations: character limits, absence of saving and possible model errors. Use the tool as an assistant, not as a replacement for pedagogical experience. The percentages example is merely an illustration; in real work always adapt recommendations to a specific student.
Open tool: Knowledge gap finder ↗Interface translations are machine-assisted and may need refinement.