arXiv:2610.33749v1 Announce Type: new Abstract: As machine translation systems approach human-level quality, traditional evaluation methodologies struggle to detect subtle translation errors. We crit...
Large Language Models (LLMs) are being tested as a cost‑effective alternative to human judgment for evaluating machine translation (MT). This study compares LLM agreement with human annotators on two MT quality schemes—Multidimensional Quality Metrics (MQM) and Error Span Annotation (ESA)—across 70 language pairs and public WMT23/25 data. Results show that while LLMs sometimes match or exceed human agreement, performance varies widely by scheme, language, and domain, and remains unreliable in many settings. The work also identifies specific challenges for both humans (e.g., fine‑grained MQM, low‑resource languages) and LLMs (e.g., minor errors, wrong language variants, span annotation).
Qwen-MT-Lumen is an open‑source translation model submitted to the WMT 2026 General Machine Translation Shared Task. It employs a four‑stage pipeline that combines continued pre‑training, supervised fine‑tuning, online preference‑optimization with an LLM judge, and on‑policy distillation from a larger teacher to a Qwen3‑14B student. The system also uses structure‑aware inference and selective repair to preserve formatting and correct decoding errors, achieving an overall mean score of 4.60 on a 1–5 rubric and matching leading proprietary systems in both translation quality and instruction adherence.
Multi‑SFT is a new, commercially permissive dataset comprising 502,818 instruction‑response pairs in five European languages, created through translation and regeneration pipelines to address gaps where simple translation is insufficient. Human evaluation shows strong inter‑annotator agreement (≥0.87) across all languages. Fine‑tuning nine models from three families on Multi‑SFT yields performance competitive with larger, proprietary mixes and official instruction‑tuned releases, especially on safety‑related benchmarks.
arXiv:2609.23490v1 Announce Type: new Abstract: Large language model (LLM) agents increasingly execute multi-step workflows through tool use and interaction with users and environments. However, curr...
arXiv:2608. 03803v1 Announce Type: cross Abstract: Multilingual language models are deployed across a hundred or more languages, yet most benchmarks test whether a model can perform a task _in_ a language rather than whether it commands the language itself, conflating fluency with proficiency.
arXiv:2512. 03086v2 Announce Type: replace-cross Abstract: Large language models (LLMs) have shown remarkable capabilities in code translation, yet their performance deteriorates in low-resource programming domains such as Fortran and emerging frameworks like CUDA, where high-quality parallel data are scarce.
The paper investigates how machine translation can be tailored to specific audiences and intents, a capability enabled by large language models (LLMs). By systematically evaluating purpose-driven MT across 50 languages, 5 model sizes, and 8 text domains, the authors find that explicit instructions significantly improve translation adaptiveness, especially for informal domains, larger models, and higher-resource languages. They also show that traditional MT metrics often penalize adapted translations and that models can self-generate useful instructions from context, closing a large portion of the adaptiveness gap.
arXiv:2610.19185v1 Announce Type: new Abstract: Evaluation of large language models (LLMs) remains largely pre-scientific: it involves many design decisions, with little guidance on how to make them....
arXiv:2606. 08840v1 Announce Type: new Abstract: Code generation models are typically compared using compact execution benchmarks and aggregate pass rates, but such summaries obscure how performance varies across programming languages, problem families, and failure modes.
arXiv:2601. 03808v2 Announce Type: replace-cross Abstract: Large language models (LLMs) have achieved notable performance in code synthesis; however, data-aware augmentation remains a limiting factor, handled via heuristic design or brute-force approaches.
The paper introduces soft‑prompt tuning as a lightweight, automated method for evaluating large language models (LLMs). By optimizing only ten soft‑prompt vectors—about 0.0006% of a 7B model’s parameters—within 150 steps, the approach adapts models to specific benchmark formats, improving format compliance while preserving knowledge accuracy. Experiments across seven models and eleven datasets show that this method matches strong evaluation‑harness baselines, enhances evaluation signals on challenging datasets, and yields cleaner outputs that align better with exact‑match scoring and independent LLM judges.