Toxic Text Classification in Portuguese: Is LLaMA 3.1 8B All You Need?
Bibliographic record
Abstract
The recognition of toxic and hate speech on social media platforms is important due to the significant risks posed to users and the digital ecosystem. Current state-of-the-art models, such as BERTimbau, have set benchmarks for Portuguese text classification, yet challenges remain in accurately detecting toxic content. This paper investigates the effectiveness of fine-tuning a smaller, open-source decoder-only model, LLaMA 3.1 8B 4bit, for this task. We propose an iterative prompt evolution method to optimize the model’s performance. Our results demonstrate that fine-tuning significantly enhances the LLaMA model’s F1-score from 0.61 to 0.75, surpassing BERTimbau in precision and matching the performance of the GPT-4o mini. However, the approach depends on the quality of the language models used for prompt evolution, highlighting the need for further research to enhance robustness in this area.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 teacher head, 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".