paper-with-me

홈 › Papers

What Language(s) Does Aya-23 Think In? How Multilinguality Affects Internal Language Representations

2025-07-27 · Katharina Trinley, Toshiki Nakai, Tatiana Anikina, Tanja Baeumel arxiv

Large language models (LLMs) excel at multilingual tasks, yet their internal language processing remains poorly understood. We analyze how Aya-23-8B, a decoder-only LLM trained on balanced multilingual data, handles code-mixed, cloze, and translation tasks compared to predominantly monolingual models like Llama 3 and Chinese-LLaMA-2. Using logit lens and neuron specialization analyses, we find: (1) Aya-23 activates typologically related language representations during translation, unlike English-centric models that rely on a single pivot language; (2) code-mixed neuron activation patterns vary with mixing rates and are shaped more by the base language than the mixed-in one; and (3) Aya-23's languagespecific neurons for code-mixed inputs concentrate in final layers, diverging from prior findings on decoder-only models. Neuron overlap analysis further shows that script similarity and typological relations impact processing across model types. These findings reveal how multilingual training shapes LLM internals and inform future cross-lingual transfer research.

📄 PDF Abstract BibTeX arXiv:2507.20279

Code (0)

등록된 구현이 없습니다.

Tasks

Cross-Lingual Transfer

Similar Papers 제목 키워드 기반

Identifying Necessary Elements for BERT's Multilinguality

2020-05-01 · Philipp Dufter, Hinrich Schütze

It has been shown that multilingual BERT (mBERT) yields high quality multilingual representations and enables effective zero-shot transfer. This is surprising given that mBERT does not use any crosslingual signal during …

Identifying Elements Essential for BERT's Multilinguality

2020-11-01 · EMNLP 2020 11 · Philipp Dufter, Hinrich Sch{\"u}tze

It has been shown that multilingual BERT (mBERT) yields high quality multilingual representations and enables effective zero-shot transfer. This is surprising given that mBERT does not use any crosslingual signal during …

What Happened in LLMs Layers when Trained for Fast vs. Slow Thinking: A Gradient Perspective

2024-10-31 · Ming Li, Yanhong Li, Tianyi Zhou

What makes a difference in the post-training of LLMs? We investigate the training patterns of different layers in large language models (LLMs), through the lens of gradient, when training with different responses and ini…

Multilinguality Does not Make Sense: Investigating Factors Behind Zero-Shot Transfer in Sense-Aware Tasks

2025-05-30 · Roksana Goworek, Haim Dubossarsky

Cross-lingual transfer allows models to perform tasks in languages unseen during training and is often assumed to benefit from increased multilinguality. In this work, we challenge this assumption in the context of two u…

Cross-Lingual TransferMultilingual NLP

Does Thought Require Sensory Grounding? From Pure Thinkers to Large Language Models

2024-08-18 · David J. Chalmers

Does the capacity to think require the capacity to sense? A lively debate on this topic runs throughout the history of philosophy and now animates discussions of artificial intelligence. I argue that in principle, there …

Philosophy