لطفا منتظر بمانید ...
0% Complete
صفحه اصلی
/
دومین همایش بین المللی هوش مصنوعی
A Systematic Review of Tokenizer Adaption Techniques for Large Language Models
نویسندگان :
Niloofar Mortazavi
1
Mostafa Amiri
2
Hesham Faili
3
1- University of Tehran
2- University of Tehran
3- University of Tehran
کلمات کلیدی :
cross-lingual adaptation،tokenizer transfer،tokenizer replacement،large language models
چکیده :
This systematic review synthesizes key research contributions in adapting Large Language Models (LLMs) to diverse linguistic contexts, focusing specifically on novel methods for transferring or replacing tokenizers and initializing associated vocabulary embeddings. The field of Large Language Model (LLM) adaptation has rapidly evolved beyond simple vocabulary expansion towards sophisticated strategies that aim for high-fidelity transfer with minimal post-hoc fine-tuning. Recent methods primarily focus on how to initialize the new embedding layer—the lexical interface—to maximize preservation of the LLM’s knowledge encoded in the frozen transformer core.
لیست مقالات
لیست مقالات بایگانی شده
Enhanced Diffusion Policy for Robust Multimodal Stock Trading through the Integration of Technical Indicators and Fundamental News
Amirali Vakili - Mahdi Shahbazi Khojasteh - Mahan Veisi - Armin Salimi-Badr
Neural-Network Surrogate Modeling for Fast Optimization of Compliant Flapping Mechanisms
Hassan Sayyaadi - Ali Zouelm
Adaptive Deep Learning Framework with Uncertainty-Aware Drift Detection, and Adversarial Robustness for Anomaly Classification in Power Systems State Estimation
Seyed Mohammad Shobeiry - Mohammad Taghi Ameli - Mohammad Sadegh Sepasian
A Master-Slave Approach for Simultaneously Controlling Two Drones when Carrying an Object
Seyyed Mohammad Ali Ardehali - Amin Faraji - Monireh Abdoos - Armin Salimi-Badr
A Thorough Analysis of How Chatbots Engage, with Aspects of Customer Experience; An In depth Review
Omid Noori
Photonic Quantum Hardware for AI Optimization Tasks
Sarah Daneshzad - Gholam-Mohammad Parsanasab
Vision-based Static Detection of Windows Malware with Rule-Augmented Feature Maps
Mahdi Seyfipoor - Mohammad Mahdi Eskandari - Siamak Mohammadi
Mamba Meets Sleep: Do State Space Models Outperform CNNs for EEG Classification?
Mostafa Mehrabi - Hamed Malek
Comparative Evaluation of YOLO-Based and CNN Models for Wheat-Head Detection in Precision Agriculture
Shahin Rostami - Shervin Danesh - Niloofar Mirzaei Chahardeh
Hybrid ANN and Ant Colony Algorithm for IoT Data Classification
Khadejeh Nemati - Safouro Ashoori - Moohamad hadi Amini
بیشتر
ثمین همایش، سامانه مدیریت همایش ها و جشنواره ها - نگارش 45.0.3