Turning Manual Tasks Into Actions: Assessing the Effectiveness of Gemini-Generated Selenium Tests
Bibliographic record
Abstract
Large Language Models (LLMs) have introduced innovative avenues for automating software testing using prompts. Despite numerous studies on software testing automation, there remains limited understanding on the effectiveness of LLM-generated Selenium tests. In this paper, we investigate the effectiveness of Gemini to produce Selenium tests from manual tasks specifications and HyperText Markup Language (HTML) code snippets. By effectiveness, we mean if the generated Selenium tests are executable and functionally accurate (meeting intended behavior specified in a manual task). To do that, we specify eight manual tasks (involving tasks related to search, filter, navigation, and form submissions) and define 25 actions for each task, using HTML code extracted from 200 web pages. These tasks require the interaction of diverse User Interface (UI) components, such as search boxes and checkboxes. The results indicate that 87.5 % of the generated Selenium tests are executable and 51.5 % of them meet the intended behavior. Manual tasks involving interaction with modals presented the greatest challenges for test generation. While carousels and buttons achieved relatively high success rates, they still accounted for many of the post-correction fixes. These components-often dynamic or context-dependent-were among those where most errors occurred during test generation.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.090 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".