The safe deployment of Large Language Models remains challenging in multilingual settings, particularly when models are exposed to adversarial or malicious prompts in underrepresented languages. In this work, we present Curupira, a Brazilian Portuguese-language guard model designed to mitigate harmful prompt exploitation. To do this, we establish a three steps methodology that involves adaptation, data generation, and fine-tuning. We also evaluate our model with two state-of-the-art open guardrail architectures. The results show that targeted fine-tuning leads to consistent improvements in safety classification for Portuguese prompts, with favorable efficiency–performance trade-offs for compact models and limited degradation in cross-lingual evaluation.

Humor is an intricate part of verbal communication and dealing with this kind of phenomenon is essential to building systems that can process language at large with all of its complexities. In this paper, we introduce Puntuguese, a new corpus of punning humor in Portuguese, motivated by previous works showing that currently available corpora for this language are still unfit for Machine Learning due to data leakage. Puntuguese comprises 4,903 manually-gathered punning one-liners in Brazilian and European Portuguese. To create negative examples that differ exclusively in terms of funniness, we carried out a micro-editing process, in which all jokes were edited by fluent Portuguese speakers to make the texts unfunny. Finally, we did some experiments on Humor Recognition, showing that Puntuguese is considerably more difficult than the previous corpus, achieving an F1-Score of 68.9%. With this new dataset, we hope to enable research not only in NLP but also in other fields that are interested in studying humor; thus, the data is publicly available.

Over the years, the review helpfulness prediction task has been the subject of several works, but remains being a challenging issue in Natural Language Processing, as results vary a lot depending on the domain, on the adopted features and on the chosen classification strategy. This paper attempts to evaluate the impact of content features and classification methods for two different domains. In particular, we run our experiments for a low resource language – Portuguese –, trying to establish a benchmark for this language. We show that simple features and classical classification methods are powerful for the task of helpfulness prediction, but are largely outperformed by a convolutional neural network-based solution.