Efficient Adaptation of LLMs for Hate Speech Detection in Low-Resource Languages: A Comparative Study on Roman Urdu

arXiv cs.AI Papers
large-language-models hate-speech-detection low-resource-languages roman-urdu peft lora comparative-study

Summary

The paper evaluates Large Language Models for hate speech detection in Roman Urdu, a low-resource language, demonstrating that Parameter-Efficient Fine-Tuning with LoRA significantly improves classification performance compared to zero-shot inference.

arXiv:2608.18142v1 Announce Type: new Abstract: It is challenging to detect hate speech in Low Resource Languages (LRLs) because of the absence of annotated data, the informality of its language structure, and the lack of standardized grammar. A good example of such a challenge is Roman Urdu which is broadly used by South Asians on social media and has a high variation while lacking contextually consistent spellings. The objective of this paper is to conduct a comprehensive assessment of Large Language Models (LLMs) for Hate Speech Detection (HSD) in Roman Urdu script and fine-tune these models using the Parameter-Efficient Fine-Tuning (PEFT) method called Low-Rank Adaptation (LoRA). To evaluate zero-shot inference, we benchmarked it against PEFT on different transformer models, including Mistral, LLaMA, Falcon, and multilingual BERT. Experiments are conducted on the PURUTT (Parallel Urdu and Roman Urdu Corpus for Toxic Comments and Transliteration) dataset with over 72,000 annotated comments. The results suggest that zero shot models perform moderately (F1 = 0.56), but updating a small fraction of the model trainable parameters improves the classification performance significantly (F1 > 0.93). Our results have shown that PEFT delivers outstanding performance alongside excellent computational efficiency, making it highly suitable for low-resource language processing tasks.
Original Article
View Cached Full Text

Cached at: 08/20/26, 10:00 AM 

# Efficient Adaptation of LLMs for Hate Speech Detection in Low-Resource Languages: A Comparative Study on Roman Urdu
Source: [https://arxiv.org/abs/2608.18142](https://arxiv.org/abs/2608.18142)
[View PDF](https://arxiv.org/pdf/2608.18142)

> Abstract:It is challenging to detect hate speech in Low Resource Languages \(LRLs\) because of the absence of annotated data, the informality of its language structure, and the lack of standardized grammar\. A good example of such a challenge is Roman Urdu which is broadly used by South Asians on social media and has a high variation while lacking contextually consistent spellings\. The objective of this paper is to conduct a comprehensive assessment of Large Language Models \(LLMs\) for Hate Speech Detection \(HSD\) in Roman Urdu script and fine\-tune these models using the Parameter\-Efficient Fine\-Tuning \(PEFT\) method called Low\-Rank Adaptation \(LoRA\)\. To evaluate zero\-shot inference, we benchmarked it against PEFT on different transformer models, including Mistral, LLaMA, Falcon, and multilingual BERT\. Experiments are conducted on the PURUTT \(Parallel Urdu and Roman Urdu Corpus for Toxic Comments and Transliteration\) dataset with over 72,000 annotated comments\. The results suggest that zero shot models perform moderately \(F1 = 0\.56\), but updating a small fraction of the model trainable parameters improves the classification performance significantly \(F1 \> 0\.93\)\. Our results have shown that PEFT delivers outstanding performance alongside excellent computational efficiency, making it highly suitable for low\-resource language processing tasks\.

## Submission history

From: Toneema Zubair \[[view email](https://arxiv.org/show-email/f0678e4b/2608.18142)\] **\[v1\]**Thu, 6 Aug 2026 11:28:19 UTC \(1,417 KB\)

Similar Articles