paper-with-me

Papers

Advancing Polish Language Modeling through Tokenizer Optimization in the Bielik v3 7B and 11B Series

2026-04-12 · Krzysztof Ociepa, Łukasz Flis, Remigiusz Kinas, Krzysztof Wróbel, Adrian Gwoździej arxiv

The development of the Bielik v3 PL series, encompassing both the 7B and 11B parameter variants, represents a significant milestone in the field of language-specific large language model (LLM) optimization. While general-purpose models often demonstrate impressive multilingual capabilities, they frequently suffer from a fundamental architectural inefficiency: the use of universal tokenizers. These tokenizers, typically designed to cover a broad spectrum of languages, often fail to capture the morphological nuances of specific languages like Polish, leading to higher fertility ratios, increased inference costs, and restricted effective context windows. This report details the transition from the universal Mistral-based tokenization to a dedicated Polish-optimized vocabulary for the Bielik v3 models, exploring the FOCUS-based embedding initialization, the multi-stage pretraining curriculum, and the subsequent post-training alignment involving Supervised Fine-Tuning, Direct Preference Optimization, and Reinforcement Learning through Group Relative Policy Optimization with verifiable rewards.

📄 PDF Abstract BibTeX arXiv:2604.10799

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning

Similar Papers 제목 키워드 기반

Cross-Family Speculative Decoding for Polish Language Models on Apple~Silicon: An Empirical Evaluation of Bielik~11B with UAG-Extended MLX-LM

2026-03-22 · Krzysztof Fonal arxiv

Speculative decoding accelerates LLM inference by using a small draft model to propose k candidate tokens for a target model to verify. While effective for same-tokenizer pairs on high-bandwidth GPUs, its applicability t…

Bielik v3 Small: Technical Report

2025-05-05 · Krzysztof Ociepa, Łukasz Flis, Remigiusz Kinas, Krzysztof Wróbel 외

We introduce Bielik v3, a series of parameter-efficient generative text models (1.5B and 4.5B) optimized for Polish language processing. These models demonstrate that smaller, well-optimized architectures can achieve per…

Language ModelingLanguage Modelling

Bielik 11B v2 Technical Report

2025-05-05 · Krzysztof Ociepa, Łukasz Flis, Krzysztof Wróbel, Adrian Gwoździej 외

We present Bielik 11B v2, a state-of-the-art language model optimized for Polish text processing. Built on the Mistral 7B v0.2 architecture and scaled to 11B parameters using depth up-scaling, this model demonstrates exc…

Language ModelingLanguage ModellingQuantization

Silver Retriever: Advancing Neural Passage Retrieval for Polish Question Answering

2023-09-15 · Piotr Rybak, Maciej Ogrodniczuk

Modern open-domain question answering systems often rely on accurate and efficient retrieval components to find passages containing the facts necessary to answer the question. Recently, neural retrievers have gained popu…

Open-Domain Question AnsweringPassage RetrievalQuestion AnsweringRetrieval

From Principles to Applications: A Comprehensive Survey of Discrete Tokenizers in Generation, Comprehension, Recommendation, and Information Retrieval

2025-02-18 · Jian Jia, Jingtong Gao, Ben Xue, Junhao Wang 외

Discrete tokenizers have emerged as indispensable components in modern machine learning systems, particularly within the context of autoregressive modeling and large language models (LLMs). These tokenizers serve as the …

Information Retrievalmultimodal generationRecommendation SystemsSurvey