Riazi-8B: An Urdu Large Language Model for Mathematical Reasoning

• URL: http://arxiv.org/abs/2606.25568v1
• Date: Wed, 24 Jun 2026 08:44:26 GMT
• Title: Riazi-8B: An Urdu Large Language Model for Mathematical Reasoning
• Abstract summary: Urdu lacks both reasoning-oriented resources and models adapted for multi-step mathematical problem solving.<n>We develop Riazi-8B, an Urdu mathematical reasoning model developed through a two-step adaptation process.<n>Our results show consistent improvements in answer correctness, reasoning quality, response completeness, and Urdu generation.
• Score: 2.1437512265005543
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Recent LLMs demonstrate strong mathematical reasoning capabilities, but existing gains rely heavily on English-centric training resources and benchmarks. As a result, reasoning performance degrades substantially in low-resource languages such as Urdu, where reasoning-oriented datasets and adapted models remain scarce. Urdu lacks both reasoning-oriented resources and models adapted for multi-step mathematical problem solving, limiting the applicability of recent progress to Urdu-speaking users. We address this gap through Riazi-8B, an Urdu mathematical reasoning model developed through a two-step adaptation process comprising continued pre-training on Urdu Wikipedia and supervised fine-tuning on Urdu Chain-of-Thought data derived from GSM8K. We evaluate Riazi-8B on MGSM-Urdu against existing Urdu instruction-tuned models. Our results show consistent improvements in answer correctness, reasoning quality, response completeness, and Urdu generation. Our findings demonstrate that combining Urdu language adaptation with reasoning-focused fine-tuning is an effective strategy for extending mathematical reasoning capabilities to low-resource languages.

Related papers

• Beyond Input Understanding: Diagnosing Multilingual Mathematical Reasoning with Directed Acyclic Trace Graphs [54.72295694649798]
  Large reasoning models (LRMs) achieve strong mathematical reasoning performance in English, but remain much less reliable in many low- and medium-resource languages.<n>We show that controlling the model's reasoning language can substantially reduce accuracy, suggesting that language also affects reasoning execution itself.<n>Experiments on the Qwen3 series across 12 languages show that non-English reasoning often suffers from reduced anchor coverage and weaker dependency fidelity.
  arXiv  Detail & Related papers  (2026-05-26T21:41:52Z)
• Crosslingual On-Policy Self-Distillation for Multilingual Reasoning [48.68444770923683]
  Crosslingual On-Policy Self-Distillation (COPSD) transfers a model's own high-resource reasoning behavior to low-resource languages.<n>Experiments on 17 low-resource African languages show that COPSD consistently improves low-resource mathematical reasoning across model sizes.
  arXiv  Detail & Related papers  (2026-05-10T14:06:09Z)
• Unlocking Reasoning Capability on Machine Translation in Large Language Models [57.60641851466707]
  Reasoning-oriented large language models (RLMs) achieve strong gains on tasks such as mathematics and coding by generating explicit intermediate reasoning.<n>We systematically evaluate several open- and closed-weights RLMs on the WMT24++ benchmark.<n>We find that enabling explicit reasoning consistently degrades translation quality across languages and models.
  arXiv  Detail & Related papers  (2026-02-16T14:05:59Z)
• UrduBench: An Urdu Reasoning Benchmark using Contextually Ensembled Translations with Human-in-the-Loop [0.17126708168238125]
  We propose a contextually ensembled translation framework with human-in-the-loop validation to develop Urdu reasoning benchmarks.<n>Using this framework, we translate widely adopted reasoning and question-answering benchmarks, including MGSM, MATH-500, CommonSenseQA, and OpenBookQA, into Urdu.<n>Our analysis reveals performance differences across (1) four datasets, (2) five task difficulty levels, (3) diverse model architectures, (4) multiple model scaling settings, and (5) language consistency tests.
  arXiv  Detail & Related papers  (2026-01-28T19:49:17Z)
• Qalb: Largest State-of-the-Art Urdu Large Language Model for 230M Speakers with Systematic Continued Pre-training [3.950299047992185]
  Urdu-a language spoken by over 230 million people-remains critically underrepresented in modern NLP systems.<n>We introduce Qalb, an Urdu language model developed through a two-stage approach: continued pre-training followed by supervised fine-tuning.<n>Our results demonstrate that continued pre-training on diverse, high-quality language data, combined with targeted instruction fine-tuning, effectively adapts foundation models to low-resource languages.
  arXiv  Detail & Related papers  (2026-01-13T02:05:05Z)
• Large Reasoning Models Are (Not Yet) Multilingual Latent Reasoners [48.68444770923683]
  Large reasoning models (LRMs) achieve strong performance on mathematical reasoning tasks.<n>LRMs often arrive at the correct answer before completing these textual reasoning steps.<n>This phenomenon has been explored in English, but its multilingual behavior remains largely unknown.
  arXiv  Detail & Related papers  (2026-01-06T13:20:17Z)
• UrduFactCheck: An Agentic Fact-Checking Framework for Urdu with Evidence Boosting and Benchmarking [43.71125919897474]
  Existing automated fact-checking systems are predominantly developed for English, leaving a significant gap for the more than 200 million Urdu speakers worldwide.<n>We present UrduFactBench and UrduFactQA, two novel benchmarks designed to enable fact-checking and factual consistency evaluation in Urdu.<n>To complement these benchmarks, we introduce UrduFactCheck, a modular fact-checking framework that incorporates both monolingual and translation-based evidence retrieval strategies.
  arXiv  Detail & Related papers  (2025-05-21T03:31:44Z)
• NoiER: An Approach for Training more Reliable Fine-TunedDownstream Task Models [54.184609286094044]
  We propose noise entropy regularisation (NoiER) as an efficient learning paradigm that solves the problem without auxiliary models and additional data. The proposed approach improved traditional OOD detection evaluation metrics by 55% on average compared to the original fine-tuned models.
  arXiv  Detail & Related papers  (2021-08-29T06:58:28Z)
• Co-occurrences using Fasttext embeddings for word similarity tasks in Urdu [0.0]
  This paper builds a corpus for Urdu by scraping and integrating data from various sources. We modify fasttext embeddings and N-Grams models to enable training them on our built corpus. We have used these trained embeddings for a word similarity task and compared the results with existing techniques.
  arXiv  Detail & Related papers  (2021-02-22T12:56:26Z)
• Efficient Urdu Caption Generation using Attention based LSTM [0.0]
  Urdu is the national language of Pakistan and also much spoken and understood in the sub-continent region of Pakistan-India. We develop an attention-based deep learning model using techniques of sequence modeling specialized for the Urdu language. We evaluate our proposed technique on this dataset and show that it can achieve a BLEU score of 0.83 in the Urdu language.
  arXiv  Detail & Related papers  (2020-08-02T17:22:33Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.