LernWerkstatt – Examples¶
Four learning units produced with LernWerkstatt. Each is a single HTML file with no external requests: it can be opened in the browser, downloaded and passed on offline.
The first two cover the same subject area but were produced from different starting positions — they show how prior knowledge and learning goal govern the treatment depth of each concept. The last two were produced from user-supplied source material rather than model knowledge.
Note
All renderer libraries are embedded and formulas are set as MathML. Once downloaded, the files work without an internet connection. Roughly 5 MB per file. The units themselves are in German.
Interpreting Statistical Results Critically – Without Calculating¶
For students of educational research who read studies but do not want to do the maths themselves. The p-value and confidence interval are derived; the test statistic remains a definition. The focus is on how robust a reported finding actually is and how to spot questionable presentations of results.
| Prerequisites | Mean, standard deviation, normal distribution, basics of significance testing |
| Sources | Model knowledge |
| Scope | 3 chapters, 8 lessons, 118 blocks |
| Glossary | 27 entries |
| Learning time | approx. 133 minutes |
Includes 14 diagrams, 9 quiz blocks, 9 prediction tasks, 8 cloze exercises, 7 flashcard sets, 6 matching exercises and one simulator.
Multiple Testing Corrections¶
The same subject area for biology doctoral researchers with routine experience in R. The basics move to the glossary; what gets derived is family-wise error rate, false discovery rate and the Benjamini-Hochberg algorithm — through to justifying the chosen correction method in a manuscript and defending it against reviewer objections.
| Prerequisites | t-test, ANOVA, linear regression, analysis in R, null hypothesis |
| Sources | Model knowledge |
| Scope | 3 chapters, 8 lessons, 85 blocks |
| Glossary | 22 entries |
| Learning time | approx. 141 minutes |
Includes 9 diagrams, 8 quiz blocks, 7 prediction tasks, 4 typeset formulas and 2 simulators.
RAG Chatbots¶
Produced from uploaded training documentation. Tokens, context windows and sampling as foundations, then building a vector knowledge base, configuring RAGflow and the typical limitations of the approach.
| Prerequisites | Basic understanding of AI, use of LLMs |
| Sources | User-supplied documents |
| Scope | 4 chapters, 12 lessons, 123 blocks |
| Glossary | 44 entries |
| Learning time | approx. 132 minutes |
This unit demonstrates the mode with own source material: terminology and claims follow the documents, are checked against them, and every concept carries a provenance note. Includes 17 diagrams, 12 quiz blocks, 10 prediction tasks, 9 tables and 2 error analyses.
AI Act Basics for University Practice¶
For staff in administration, research and teaching. Covers when a working tool counts as an AI system under the regulation, how to classify one's own role as deployer, user or third party, and what consequences follow from the four risk tiers — through to a review scheme for everyday work and the question of when a planned deployment needs to be escalated.
| Prerequisites | Digital work processes, own role in the university context, familiarity with AI applications, compliance and data protection basics |
| Sources | User-supplied documents |
| Scope | 3 chapters, 9 lessons, 84 blocks |
| Glossary | 27 entries |
| Learning time | approx. 107 minutes |
An example of regulatory content where binding claims to the sources matters particularly: statements are checked against the supplied documents. Includes 16 diagrams, 9 tables, 9 quiz blocks, 8 prediction tasks, 7 matching exercises and 6 flashcard sets.