Are Multilingual Models Actually Improving? Revisiting Cross-Lingual Transfer
cs.CL
Submitted: 2026-06-20
Updated: 2026-11-09
Terminology
Sources
- InfoXLM: An Information-Theoretic Framework for Cross-Lingual Language Model Pre-Training
- Training Verifiers to Solve Math Word Problems
- Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
- Training on the Test Task Confounds Evaluation and Emergence
- ECLeKTic: a Novel Challenge Set for Evaluation of Cross-Lingual Knowledge Transfer
- A Survey on LLM-as-a-Judge
- SimpleQA Verified: A Reliable Factuality Benchmark to Measure Parametric Knowledge
- MuBench: Assessment of Multilingual Capabilities of Large Language Models Across 61 Languages
- Multilingual Needle in a Haystack: Investigating Long-Context Behavior of Multilingual Large Language Models
- XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization
- Large Language Models Are Cross-Lingual Knowledge-Free Reasoners
- Quantifying Language Disparities in Multilingual Large Language Models
- Beneath the Surface of Consistency: Exploring Cross-lingual Knowledge Representation Sharing in LLMs
- Domain Adaptation: Learning Bounds and Algorithms
- Accuracy on the Line: On the Strong Correlation Between Out-of-Distribution and In-Distribution Generalization
- gpt-oss-120b & gpt-oss-20b Model Card
- Mind the Gap... or Not? How Translation Errors and Evaluation Details Skew Multilingual Results
- Rethinking Cross-lingual Gaps from a Statistical Viewpoint
- ImageNot: A contrast with ImageNet preserves model rankings
- Language Models are Multilingual Chain-of-Thought Reasoners
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data