paper
When English Isn't the Best Teacher: Source Language Effects in Cross-Lingual In-Context Learning
paperactiveprovisional
when-english-isn-t-the-best-teacher-source-language-effects-in-cross-lingual-in-context-learning-60020e86·1 events·first seen 4h agoAliases: When English Isn't the Best Teacher: Source Language Effects in Cross-Lingual In-Context Learning
More like this (12)
Reinforcement Learning Elicits Contextual Learning of Unseen Language Translationin-context learningIn-Context Multiple Instance Learningcontrastive learningThe Shibboleth Effect: Auditing the Cross-Lingual Distributional Skew of Large Language ModelsUnintended Effects of Geographic Conditioning in Large Language ModelsLatent Context Language ModelsOn The Effectiveness-Fluency Trade-Off In LLM Conditioning: A Systematic StudyRevisiting the Systematicity in Negation in the Era of In-Context LearningWhen Does Mixing Help? Analyzing Query Embedding Interpolation in Multilingual Dense RetrievalBeyond Accuracy: Community Perspectives on Machine TranslationReasoning over Grammar: Can Synthetic Linguistic Reasoning Traces Enhance Low-Resource Machine Translation?
Recent events (1)
Cross-lingual in-context learning source language selection challenges fine-tuning assumptions
A new arXiv paper conducts a broad empirical study of cross-lingual transfer in few-shot in-context learning (ICL), spanning seven tasks, six models, and a typologically diverse set of languages. The study finds that conventional heuristics from supervised fine-tuning — such as relying on linguistic similarity or data availability — do not consistently transfer to the ICL regime. The authors also analyze language confusion as a key obstacle in generative cross-lingual ICL and propose alternative heuristics for source language selection.