تقرير
On the Origins of Linear Representations in Large Language Models
| العنوان: | On the Origins of Linear Representations in Large Language Models |
|---|---|
| المؤلفون: | Jiang, Yibo, Rajendran, Goutham, Ravikumar, Pradeep, Aragam, Bryon, Veitch, Victor |
| سنة النشر: | 2024 |
| المجموعة: | Computer Science Statistics |
| مصطلحات موضوعية: | Computer Science - Computation and Language, Computer Science - Machine Learning, Statistics - Machine Learning |
| الوصف: | Recent works have argued that high-level semantic concepts are encoded "linearly" in the representation space of large language models. In this work, we study the origins of such linear representations. To that end, we introduce a simple latent variable model to abstract and formalize the concept dynamics of the next token prediction. We use this formalism to show that the next token prediction objective (softmax with cross-entropy) and the implicit bias of gradient descent together promote the linear representation of concepts. Experiments show that linear representations emerge when learning from data matching the latent variable model, confirming that this simple structure already suffices to yield linear representations. We additionally confirm some predictions of the theory using the LLaMA-2 large language model, giving evidence that the simplified model yields generalizable insights. |
| نوع الوثيقة: | Working Paper |
| URL الوصول: | http://arxiv.org/abs/2403.03867 |
| رقم الأكسشن: | edsarx.2403.03867 |
| قاعدة البيانات: | arXiv |
| الوصف غير متاح. |