Cross-lingual Few-shot Learning for Persian Sentiment Analysis with Incremental Adaptation

This research examines cross-lingual sentiment analysis using few-shot learning and incremental learning methods in Persian. The main objective is to develop a model capable of performing sentiment analysis in Persian using limited data, while getting prior knowledge from high-resource languages. To achieve this, three pre-trained multilingual models (XLM-RoBERTa, mDeBERTa, and DistilBERT) were employed, which were fine-tuned using few-shot and incremental learning approaches on small samples of Persian data from diverse sources, including X, Instagram, Digikala, Snappfood, and Taaghche. This variety enabled the models to learn from a broad range of contexts. Experimental results show that the mDeBERTa and XLM-RoBERTa achieved high performances, reaching 96% accuracy on Persian sentiment analysis. These findings highlight the effectiveness of combining few-shot learning and incremental learning with multilingual pre-trained models.

The Cross-Lingual Cost: Retrieval Biases in RAG over Arabic-English Corpora

Cross-lingual retrieval-augmented generation (RAG) is a critical capability for retrieving and generating answers across languages. Prior work in this context has mostly focused on generation and relied on benchmarks derived from open-domain sources, most notably Wikipedia. In such settings, retrieval challenges often remain hidden due to language imbalances, overlap with pretraining data, and memorized content. To address this gap, we study Arabic-English RAG in a domain-specific setting using benchmarks derived from real-world corporate datasets. Our benchmarks include all combinations of languages for the user query and the supporting document, drawn independently and uniformly at random. This enables a systematic study of multilingual retrieval behavior. Our findings reveal that retrieval is a critical bottleneck in cross-lingual domain-specific scenarios, with significant performance drops occurring when the user query and supporting document languages differ. A key insight is that these failures stem primarily from the retriever's difficulty in ranking documents across languages. Finally, we propose a simple retrieval strategy that addresses this source of failure by enforcing equal retrieval from both languages, resulting in substantial improvements in cross-lingual and overall performance. These results highlight meaningful opportunities for improving multilingual retrieval, particularly in practical, real-world RAG applications.

Conditional Unigram Tokenization with Parallel Data

We introduce conditional unigram tokenization, a novel approach that extends unigram tokenization by conditioning target token probabilities on source-language tokens from parallel data. Given a fixed source tokenizer, our method learns a target tokenizer that maximizes cross-lingual semantic alignment. We evaluate our tokenizer on four language pairs across different families and resource levels, examining intrinsic properties and downstream performance on machine translation and language modeling. While our conditional tokenizer maintains comparable statistical properties to standard unigram tokenizers, results are mixed: we observe no improvements in machine translation quality, but find consistent perplexity reductions in language modeling. We hypothesize that quadratic scaling of conditional probability estimation with respect to the vocabulary size creates a data efficiency bottleneck. Our findings suggest that alternative parameterizations may be necessary for practical cross-lingual tokenization.

StreamUni: Achieving Streaming Speech Translation with a Unified Large Speech-Language Model

Streaming speech translation (StreamST) requires determining appropriate timing, known as policy, to generate translations while continuously receiving source speech inputs, balancing low latency with high translation quality. However, existing StreamST methods typically operate on sentence-level speech segments, referred to as simultaneous speech translation (SimulST). In practice, they require collaboration with segmentation models to accomplish StreamST, where the truncated speech segments constrain SimulST models to make policy decisions and generate translations based on limited contextual information. Moreover, SimulST models struggle to learn effective policies due to the complexity of speech inputs and cross-lingual generation. To address these challenges, we propose StreamUni, which achieves StreamST through a unified Large Speech-Language Model (LSLM). Specifically, StreamUni incorporates speech Chain-of-Thought (CoT) in guiding the LSLM to generate multi-stage outputs. Leveraging these multi-stage outputs, StreamUni simultaneously accomplishes speech segmentation, policy decision, and translation generation, completing StreamST without requiring massive policy-specific training. Additionally, we propose a streaming CoT training method that enhances low-latency policy decisions and generation capabilities using limited CoT data. Experiments demonstrate that our approach achieves state-of-the-art performance on StreamST tasks.

MuRating: A High Quality Data Selecting Approach to Multilingual Large Language Model Pretraining

Data quality is a critical driver of large language model performance, yet existing model-based selection methods focus almost exclusively on English. We introduce MuRating, a scalable framework that transfers high-quality English data-quality signals into a single rater for 17 target languages. MuRating aggregates multiple English "raters" via pairwise comparisons to learn unified document-quality scores,then projects these judgments through translation to train a multilingual evaluator on monolingual, cross-lingual, and parallel text pairs. Applied to web data, MuRating selects balanced subsets of English and multilingual content to pretrain a 1.2 B-parameter LLaMA model. Compared to strong baselines, including QuRater, AskLLM, DCLM and so on, our approach boosts average accuracy on both English benchmarks and multilingual evaluations, with especially large gains on knowledge-intensive tasks. We further analyze translation fidelity, selection biases, and underrepresentation of narrative material, outlining directions for future work.

Text2Cypher Across Languages: Evaluating Foundational Models Beyond English

Recent advances in large language models have enabled natural language interfaces that translate user questions into database queries, such as Text2SQL, Text2SPARQL, and Text2Cypher. While these interfaces enhance database accessibility, most research today focuses solely on English, with limited evaluation in other languages. This paper investigates the performance of foundational LLMs on the Text2Cypher task across multiple languages. We create and release a multilingual test set by translating English questions into Spanish and Turkish while preserving the original Cypher queries, enabling fair cross-lingual comparison. We evaluate multiple foundational models using standardized prompts and metrics. Our results show a consistent performance pattern: highest on English, then Spanish, and lowest on Turkish. We attribute this to differences in training data availability and linguistic characteristics. Additionally, we explore the impact of translating task prompts into Spanish and Turkish. Results show little to no change in evaluation metrics, suggesting prompt translation has minor impact. Our findings highlight the need for more inclusive evaluation and development in multilingual query generation. Future work includes schema localization and fine-tuning across diverse languages.

Multi-lingual Functional Evaluation for Large Language Models

Multi-lingual competence in large language models is often evaluated via static data benchmarks such as Belebele, M-MMLU and M-GSM. However, these evaluations often fail to provide an adequate understanding of the practical performance and robustness of models across multi-lingual settings. In response, we create multi-lingual functional benchmarks -- Cross-Lingual Grade School Math Symbolic (CL-GSM Symbolic) and Cross-Lingual Instruction-Following Eval (CL-IFEval)-- by translating existing functional benchmark templates from English to five additional languages that span the range of resources available for NLP: French, Spanish, Hindi, Arabic and Yoruba. Our results reveal that some static multi-lingual benchmarks capture functional performance much more closely than others (i.e. across models, there is a 24%, 17% and 18% decrease in performance between M-GSM and CL-GSM Symbolic in English, French and Spanish respectively; similarly there's a 15 - 24% performance drop across languages between Belebele and CL-IFEval, and only a 0.5% to 3% performance drop between M-MMLU and CL-IFEval). Similarly, we find that model robustness across languages varies significantly, with certain languages (eg. Arabic, English) being the most consistently well performing across evaluation iterations.

MuBench: Assessment of Multilingual Capabilities of Large Language Models Across 61 Languages

Multilingual large language models (LLMs) are advancing rapidly, with new models frequently claiming support for an increasing number of languages. However, existing evaluation datasets are limited and lack cross-lingual alignment, leaving assessments of multilingual capabilities fragmented in both language and skill coverage. To address this, we introduce MuBench, a benchmark covering 61 languages and evaluating a broad range of capabilities. We evaluate several state-of-the-art multilingual LLMs and find notable gaps between claimed and actual language coverage, particularly a persistent performance disparity between English and low-resource languages. Leveraging MuBench's alignment, we propose Multilingual Consistency (MLC) as a complementary metric to accuracy for analyzing performance bottlenecks and guiding model improvement. Finally, we pretrain a suite of 1.2B-parameter models on English and Chinese with 500B tokens, varying language ratios and parallel data proportions to investigate cross-lingual transfer dynamics.

Prompt, Translate, Fine-Tune, Re-Initialize, or Instruction-Tune? Adapting LLMs for In-Context Learning in Low-Resource Languages

LLMs are typically trained in high-resource languages, and tasks in lower-resourced languages tend to underperform the higher-resource language counterparts for in-context learning. Despite the large body of work on prompting settings, it is still unclear how LLMs should be adapted cross-lingually specifically for in-context learning in the low-resource target languages. We perform a comprehensive study spanning five diverse target languages, three base LLMs, and seven downstream tasks spanning over 4,100 GPU training hours (9,900+ TFLOPs) across various adaptation techniques: few-shot prompting, translate-test, fine-tuning, embedding re-initialization, and instruction fine-tuning. Our results show that the few-shot prompting and translate-test settings tend to heavily outperform the gradient-based adaptation methods. To better understand this discrepancy, we design a novel metric, Valid Output Recall (VOR), and analyze model outputs to empirically attribute the degradation of these trained models to catastrophic forgetting. To the extent of our knowledge, this is the largest study done on in-context learning for low-resource languages with respect to train compute and number of adaptation techniques considered. We make all our datasets and trained models available for public use.

Targeted Lexical Injection: Unlocking Latent Cross-Lingual Alignment in Lugha-Llama via Early-Layer LoRA Fine-Tuning

Large Language Models (LLMs) have demonstrated remarkable capabilities, yet their performance in low-resource languages (LRLs), such as Swahili, often lags due to data scarcity and underrepresentation in pre-training. A key challenge is achieving robust cross-lingual lexical alignment, crucial for tasks like translation and cross-lingual information retrieval. This paper introduces Targeted Lexical Injection (TLI), a novel and efficient fine-tuning approach. We first demonstrate that Lugha-Llama-8B-wura, a Swahili-centric LLM, exhibits strong, near-perfect lexical alignment for Swahili-English word pairs in its early internal layers (specifically Layer 2, with ~0.99998 average cosine similarity based on a pilot study), a capability not fully reflected in its final output representations (baseline ~0.32 similarity on our evaluation set). TLI leverages this insight by using Low-Rank Adaptation (LoRA) and a contrastive learning objective to fine-tune the model, specifically targeting embeddings from this empirically identified optimal early layer. Our experiments show that TLI significantly improves the output-level lexical alignment for 623 trained Swahili-English word pairs, increasing average cosine similarity from 0.3211 to 0.4113 (+28.08%, p < 1.33 x 10^-240). More importantly, these improvements generalize remarkably well to 63 unseen control word pairs, with similarity increasing from 0.3143 to 0.4033 (+28.32%, p < 7.17 x 10^-27). These findings suggest TLI enhances the model's ability to preserve and propagate its inherent early-layer cross-lingual knowledge, offering a parameter-efficient and effective strategy for improving lexical alignment in LRL-focused LLMs.