arXiv subjects

Fajri Koto

Publications and source records attributed to Fajri Koto.

At least 19 recordsLinked to original sources

arXiv2026

Steering with Multilingual Lexicons for Sentiment Control in Low-Resource Languages

Activation steering offers a lightweight mechanism for controlling model behavior, but existing methods depend on sentence-level contrastive data that is unavailable for most of the world's languages. We show that multilingual sentiment lexicons are a viable substitute across three model families: aggregated multilingual lexicons outperform English-derived steering vectors for mid- and low-resource languages, while monolingual lexicons remain stronger for languages a model has specialized in during pre-training. The steering efficacy is mechanistically dictated by both representational density and pre-training priors. Large-magnitude English vectors overwhelm the compact representations of low-resource languages, causing "linguistic collapse." This fragility is mitigated by the norm-dampening of multilingual aggregation. Architectural biases, including bilingual specialization penalty and latent numerical rigidity, also heavily constrain cross-lingual steering responsiveness. Ultimately, our work establishes lexicons as a practical steering signal and offers a comprehensive mechanistic account of cross-lingual steering dynamics.

cs.CL↗

arXiv2026

Text Generation in Low-Resource Languages Using Human-Curated Vocabularies

Synthetic text generation offers a promising approach to alleviating data scarcity in low-resource languages. However, its effectiveness is inherently limited for languages that are scarcely represented, or entirely absent, in an LLM's pretraining data. At the same time, many such languages have human-curated vocabularies that provide rich lexical and semantic information, often accompanied by grammatical information such as part-of-speech labels. In this work, we ask whether LLMs can leverage these resources to generate coherent text in extremely low-resource and unseen languages. We introduce a framework that combines human-curated vocabularies with a small number of human-written exemplars that provide syntactic signals, while sampling candidate words according to a Zipfian distribution. Across languages with varying levels of resource availability, we evaluate the generated text using lexical diversity, off-target ratio, and human judgments of fluency. Our results show that reasoning-enabled LLMs can generate fluent text even in extremely low-resource languages, with particularly strong results at the sentence level and promising performance at the paragraph level. Our findings highlight a practical pathway for bootstrapping synthetic data in low-resource languages, expanding the amount of usable text even when existing digital resources are extremely limited. Our source code is available at https://github.com/michaelsinanta/vocab2text.

cs.CL↗

arXiv2026

Aligning Cultural Behaviors in Arabic Language Models via Preferences

Arabic LLMs have improved rapidly, but aligning them with country-specific cultural behavioral preferences remains underexplored. Previous work has focused primarily on instruction-tuning for factual knowledge, overlooking nuanced cultural behaviors. We propose ArabPref, which to our knowledge is the first large-scale pan-Arab preference data set for the alignment of cultural behavior conditioned on the country. Our contribution includes two resources in English and Modern Standard Arabic: a culturally grounded preference dataset for training and evaluation, and a multiple-choice benchmark designed to assess culturally aligned behavior across nations. All test data are validated by native speakers to ensure authenticity. To optimize cultural alignment, we experiment with post-training methods across multilingual and Arabic-centric language models, evaluating performance on both generation tasks and our cultural multiple-choice benchmark. By releasing an evaluation benchmark, we provide a foundation for systematically studying and improving culturally aligned behavior in language models.

cs.CL↗

arXiv2026

Simulating Student-Tutor Interactions with LLMs for Multilingual Math Feedback

Large language models (LLMs) have demonstrated the ability to generate formative feedback and instructional hints in English, making them increasingly relevant for AI-assisted education. However, their ability to provide effective instructional support across different languages, especially for mathematically grounded reasoning tasks, remains largely unexamined. In this work, we present the first large-scale simulation of multilingual student-tutor interactions using LLMs. A stronger model plays the role of the tutor, generating feedback in the form of hints, while a weaker model simulates the student. We explore 352 experimental settings across 11 typologically diverse languages, four state-of-the-art LLMs, and multiple prompting strategies to assess whether language-specific feedback leads to measurable learning gains. Our study examines how student input language, tutor feedback language, model choice, and language resource level jointly influence performance. Results show that English hints are generally preferred, even by multilingual student LLMs, with few exceptions that lead to improved student learning outcomes. These findings offer practical insights for developing multilingual, LLM-based educational tools that are both effective and inclusive.

cs.CL↗

arXiv2026

AgentFly: Scaling Agentic Reinforcement Learning with Unified Resource System

Methods to build LLM agents have evolved from prompt engineering and supervised finetuning to agentic reinforcement learning (agentic RL). However, agentic RL remains bottlenecked by its surrounding systems: agents must interact with heterogeneous environments, such as sandboxes, model services, and external APIs. Their allocation, reuse, and lifecycle dominate rollout cost and cap the scale at which training becomes practical. In this work, we present AgentFly, an agentic RL framework built with a unified resource layer that treats each of these environments as a distinct, typed resource scheduled through one engine, with per-tool acquisition for multi-turn reuse, asynchronous backpressure, and rollout versus global-scoped lifecycles. AgentFly adopts a four-layer design: (I) agent layer that abstracts the agent, tool, and reward concepts, decomposing agentic RL into defining agents, tools, and reward functions; (II) rollout layer that composes these into agent loops and computes rewards; (III) context layer that organizes rollouts, injects contextual information, and arranges resources; and (IV) a low-level resource layer that performs resource management. We provide a suite of prebuilt tools and environments, demonstrate successful agent training across multiple tasks and models, and report the first controlled cross-framework throughput comparison against agentic RL frameworks.

cs.AI↗

arXiv2026

Low-Resource Safety Failures Are Action Failures, Not Representation Failures

Language models often answer harmful requests in low-resource languages (LRLs) that they refuse in high-resource languages (HRLs). Across three instruction-tuned models and 23 languages, harmful refusal falls from 87.9% in HRLs to 43.9% in LRLs, while harmless refusal remains low. A common explanation is that models represent harmfulness weakly in LRLs. We test whether harmfulness is instead represented but does not reliably produce refusal. Across three models, a harmfulness direction learned from HRL activations still separates harmful from harmless LRL prompts, showing that complete absence of harmfulness information cannot explain many failures. However, harmfulness scores shift downward for LRL prompts, making harmful prompts less likely to reach the range associated with refusal. Motivated by this shift, we train a low-rank logistic classifier on HRL activations and calibrate its threshold with a few target-language examples. During generation, the classifier conditionally adds or ablates the HRL harmfulness direction. With the same HRL data and 32 target-language examples per class, CAST remains limited by low harmful refusal and AdaSteer by high harmless refusal, yielding mean refusal selectivity ($Δ$ = harmful - harmless refusal) of 33.6 and 6.8, respectively. Our intervention reaches 54.5 while preserving MMLU utility. HRL-only calibration improves selectivity for Qwen and Gemma, whereas Llama benefits from target-language calibration. These results show that recalibrating existing representations can offer a training-free method for repairing low-resource safety failures.

cs.CL↗

arXiv2026

SkillGym: Training Skill-Use Agents with Automatic Verifiable Environment Generation

Skills equip LLM agents with professional knowledge and guidance to complete long-horizon and complex tasks. Although skills have been widely adopted in recent agent paradigms and harnesses, how to synthesize reliable training data and how to train agents for skill use remain underexplored. In this work, we propose SkillGym, an automatic pipeline to build verifiable environments, collect trajectories, and train skill-use agents. SkillGym first crawls a large volume of skills from the internet, then keeps those whose workflows can run reproducibly offline. A builder-reviewer pipeline is used to construct difficulty-controlled tasks, spanning four task types, each with a reference solution and an executable verifier. With this pipeline, we build 6.8k environments and collect 19k verified successful trajectories for supervised finetuning. Finetuning on these trajectories improves LLMs of different families and sizes, from 2B to 122B parameters across four skill-use benchmarks; Our Qwen3.5-9B SFT model outperforms the 397B untrained model on two of them. Further analysis shows that training teaches agents to invoke skills, raising the rate of reading the relevant skill from 28% to 96%, and that the gains hold across reasoning structures, extending to task types that form a minority of the training data and to skills held out from training

cs.AI↗

arXiv2026

Parallel Tokenizers: Rethinking Vocabulary Design in Cross-Lingual Transfer of Low-Resource Languages

Tokenization forms the basis of multilingual language models, yet existing methods often limit cross-lingual transfer by mapping semantically equivalent words to different embeddings. For example, 'I eat rice' in English and 'Ina cin shinkafa' in Hausa are typically mapped to different vocabulary indices, preventing shared representations and limiting cross-lingual generalization. This problem is even more pronounced in low-resource languages, where shared representations could offer the greatest benefit. We introduce parallel tokenizers, a new framework that first trains tokenizers monolingually and then aligns their vocabularies exhaustively using bilingual dictionaries or word-to-word translation. This alignment enforces a shared semantic space across languages while naturally improving fertility balance. To assess their effectiveness, we pretrain a transformer encoder from scratch on thirteen low-resource languages and evaluate it on sentiment analysis, hate speech detection, emotion classification, and sentence embedding similarity. Across all tasks, models trained with parallel tokenizers outperform conventional multilingual baselines, confirming that rethinking tokenization is essential for advancing multilingual representation learning--especially in low-resource settings.

cs.CL↗

arXiv2026

Sci-Rho: A Multilingual Visually-Grounded Symbolic Benchmark for STEM Problems

Symbolic benchmarks have emerged as a key approach to assess model robustness under minor modifications to STEM-related questions. However, existing symbolic benchmarks mostly remain limited to mathematical reasoning, lack visual grounding, and are predominantly in English. In this work, we introduce Sci-Rho (Science Rhobustness), a dynamic benchmark for visually-grounded STEM problems spanning five subjects and seven languages, comprising 4,242 problem templates (606 per language) crafted by domain experts, including Olympiad medalists. Each template is implemented as executable Python code that generates diverse but equivalent problem instances by varying numerical values, visual patterns, geometric shapes, color schemes, and function types, resulting in 42,420 instances in total, each paired with reasoning steps and ground-truth solutions. We evaluated 17 state-of-the-art VLMs and discovered a noticeable gap between worst-case accuracy (defined as the proportion of problem templates that a model answers correctly across every generated variation) and average accuracy. We also discovered that smaller models show noticeable performance degradation across languages, whereas proprietary and larger models remain robust. Step-level evaluation reflects this same trend, revealing a significant gap between average F1 and worst-case F1 scores. Finally, our inspection of attention heads of a VLM reveals substantial cross-lingual variation in the relative attention allocated to image tokens compared to text tokens. Our work highlights the importance of evaluation beyond static benchmarks as a metric to measure the quality of VLMs.

cs.CV↗

arXiv2026

TextAlign: Preference Alignment for Text Rendering with Hierarchical Rewards

Faithful text rendering remains a persistent weakness of large text-to-image generative models, as it requires both semantic instruction following and fine-grained glyph-level structure. Prior methods often improve this ability through architecture-specific modules or encoder modifications, which complicate deployment across foundation models. We study text rendering as a post-training preference-alignment problem and propose TextAlign, a non-invasive framework that keeps the generator architecture unchanged. The key component is a hierarchical vision-language model (VLM)-based reward that decomposes rendering errors into global, word, and glyph levels, then converts binary defect judgments into a scalar preference signal. The resulting signal supports both Group Relative Policy Optimization (GRPO) and Direct Preference Optimization (DPO). Experiments on FLUX.1-dev and Z-Image-Turbo show consistent gains in OCR-based text accuracy without degrading general generation quality. Compared with strong foundation and text-rendering baselines, including SD3.5, Qwen-Image, AnyText, and TextDiffuser, these results indicate that reward design offers a scalable alternative to model redesign for improving text rendering.

cs.CV↗

arXiv2026

Right Frame, Wrong Rule: Cultural Cues Expose the Financial Knowledge Gap They Were Meant to Close

When a question has valid answers under different normative frameworks, a language model must decide which framework to use and whether it can answer correctly within it. We call this setting normative pluralism and study it in Islamic finance using a four-choice taxonomy that separates framework selection from within-framework correctness. This separation reveals the stereotype trap: a cultural cue steers a model toward one framework, but the model selects an incorrect answer within that framework. Across twelve models, two languages, and fifty demographic signals, cultural cues change framework selection and reveal substantial differences in accuracy, especially among non-frontier models. Under the strongest signal, large open-weight models select the Islamic framework 97% of the time. A two-choice evaluation would report near-perfect alignment, although 57--66% of those selections are incorrect. These findings motivate, but do not directly test, the competence-conditioned routing hypothesis: models may favor frameworks where they are more accurate, while cultural cues may expose framework-specific competence gaps.

cs.CL↗

arXiv2026

Training and Benchmarking Code Generation for Physics-Inspired Animations

Large language models (LLMs) have been widely studied in areas such as mathematical reasoning, complex coding, and scientific problem solving. However, their ability to generate executable code that visually depicts physical scenarios and their qualitative dynamics remains underexplored. We propose SimuScene, the first systematic study that trains and evaluates LLMs on code generation for physics-inspired animations across 52 concepts spanning five physics domains. We build an automated data collection pipeline with human verification to ensure data quality. The resulting dataset contains 7,659 scenarios, including a 334-example human-verified test set. We evaluate 10 contemporary LLMs and find that even the strongest model achieves only a 21.5\% Avg@8 accuracy, demonstrating the difficulty of generating animations that are both executable and visually aligned with physical scenario descriptions. Finally, we introduce a reinforcement learning pipeline that uses visual rewards from code-generated videos to train text-only LLMs, with a vision-language model evaluating videos through verification questions. Experiments show that training with our data and video-based rewards improves LLM performance on physics-inspired animation generation.

cs.LG↗

arXiv2026

CultureTalk-ID: A Multi-Task Dialogue Benchmark for Cultural Commonsense in Indonesian Local Languages

Culture is lived through conversation, yet existing Indonesian cultural commonsense benchmarks evaluate LLMs on short and isolated prompts, stripping away the dialogic context in which cultural nuances actually surface. We introduce CultureTalk-ID, the first dialogue-based benchmark for cultural commonsense in Indonesian and its local languages, comprising 4,496 culturally grounded dialogues across 11 languages and 13 culturally salient topics, curated through a multi-stage human pipeline with native speakers to ensure authenticity. CultureTalk-ID introduces three complementary tasks, namely dialogue-based multiple-choice cultural commonsense reasoning, culturally faithful machine translation, and language steering, which jointly probe whether LLMs can understand, transfer, and generate culturally grounded language.

cs.CL↗

arXiv2026

Jais 2: A Family of Arabic-Centric Open Large Language Models

Jais 2 is a family of Arabic-centric large language models developed jointly by MBZUAI, Cerebras, and Inception, designed to advance Arabic-centric language modeling, with strong performance across the Arabic and culturally grounded benchmarks evaluated in this report. The family includes, to our knowledge, the largest open Arabic-centric LLM trained from scratch at 70B parameters, and a competitive 8B-parameter variant among the evaluated open models. A custom Arabic-centric vocabulary enables efficient training and inference. In addition, an optimized architecture and training recipe yield highly compute-efficient training. With a substantially smaller token budget than comparable models, Jais 2 achieves strong Arabic performance on the benchmarks considered in this report and competitive English results. The models obtain leading results among the evaluated open models on OALL2 and AraGen. They also perform strongly on several culturally grounded Arabic benchmarks, including poetry, religion, cuisine, and dream interpretation, as well as in general tasks such as translation and summarization. We release the models in HuggingFace under a commercially permissive license. Jais 2 70B is also released as a chat app on the Web, iOS, and Android; it runs on Cerebras hardware, delivering up to 2,000 tokens per second, and enabling high-throughput Arabic-centric chat serving in our deployment setting. By uniting scale, linguistic diversity, cultural fidelity, openness, and speed, Jais 2 provides an open-weight foundation intended to support further research and development in Arabic-centric LLMs.

cs.CL↗

arXiv2026

Preserving Fairness and Safety in Quantized LLMs Through Critical Weight Protection

Quantization is widely adopted to reduce the computational cost of large language models (LLMs); however, its implications for fairness and safety, particularly in dynamic quantization and multilingual contexts, remain underexplored. In this work, we conduct a systematic study of how static and dynamic quantization methods impact fairness and safety across benchmarks measuring intrinsic and extrinsic bias and safety alignment. For fairness, we evaluate English, French, Dutch, Spanish, and Turkish; for safety, we focus on English, Korean, and Arabic. Our findings reveal that quantization consistently degrades fairness and safety, with dynamic methods demonstrating greater stability than static ones. Moreover, fairness degradation varies across languages, while safety deterioration is especially pronounced in non-English settings. To address these risks, we introduce Critical Weight Protection, a novel technique that identifies and preserves fairness- and safety-critical weights during quantization. This approach effectively mitigates bias and safety deterioration without costly retraining or alignment, maintaining trustworthiness while retaining efficiency.

cs.CL↗

arXiv2026

Exploring Language-Agnosticity in Function Vectors: A Case Study in Machine Translation

Function vectors (FVs) are vector representations of tasks extracted from model activations during in-context learning. While prior work has shown that multilingual model representations can be language-agnostic, it remains unclear whether the same holds for function vectors. We study whether FVs exhibit language-agnosticity, using machine translation as a case study. Across three decoder-only multilingual LLMs, we find that translation FVs extracted from a single English$\to$X direction transfer to other target languages, consistently improving the rank of correct translation tokens across multiple unseen languages. We further find that the highest-gain tokens span multiple languages and that translation FVs across directions share most of their top-ranked heads, indicating that the FV encodes a largely language-agnostic translation signal rather than a language-pair-specific mapping.

cs.CL↗

arXiv2026

Revisiting Metric Reliability for Fine-grained Evaluation of Machine Translation and Summarization in Indian Languages

While automatic metrics drive progress in Machine Translation (MT) and Text Summarization (TS), existing metrics have been developed and validated almost exclusively for English and other high-resource languages. This narrow focus leaves Indian languages, spoken by over 1.5 billion people, largely overlooked, casting doubt on the universality of current evaluation practices. To address this gap, we introduce ITEM, a large-scale benchmark that systematically evaluates the alignment of 29 automatic metrics with human judgments across six major Indian languages, enriched with fine-grained annotations. Our extensive evaluation, covering agreement with human judgments, sensitivity to outliers, language-specific reliability, inter-metric correlations, and resilience to controlled perturbations reveals four central findings: (1) LLM-based evaluators show the strongest alignment with human judgments at both segment and system levels; (2) outliers exert a significant impact on metric-human agreement; (3) In TS, metrics are more effective at capturing content fidelity, whereas in MT, they better reflect fluency; and (4) Metrics differ in their robustness and sensitivity when subjected to diverse perturbations. Collectively, these findings offer critical guidance for advancing metric design and evaluation in Indian languages.

cs.CL↗

arXiv2026

Multilingual Idioms in Sentences and Conversations Across High-, Medium-, and Low-Resource Languages

Idiomatic expressions pose a major challenge for multilingual NLP because their meanings shift between figurative and literal usage, often requiring context for accurate interpretation. Prior work has focused on high-resource languages typically evaluates isolated idiom-meaning questions, overlooking realistic discourse. We introduce MIDI, a multilingual idiom dataset spanning 3 high-, 3 medium-, and 12 low-resource languages, curated by native speakers. Unlike previous datasets, MIDI provides idioms embedded in both sentence-level and conversational contexts, capturing both literal and figurative readings. Benchmarking state-of-the-art models shows that idiom comprehension degrades in low-resource languages and that, in all resource tiers, literal interpretations are substantially harder than figurative ones. Conversational context improves performance but does not eliminate these disparities. Through controlled tests and interventions on hidden representations, we further separate memorization from reasoning, exposing core limitations of current models.

cs.CL↗