Bielik 7B v0.1: A Polish Language Model -- Development, Insights, and Evaluation

TL;DR

Bielik 7B v0.1 is a new 7-billion-parameter Polish language model that introduces innovative training techniques and evaluation frameworks, achieving significant performance improvements and setting new benchmarks in Polish NLP tasks.

Contribution

The paper presents Bielik 7B v0.1, a novel Polish language model with innovative training methods and comprehensive evaluation frameworks, advancing NLP capabilities for Polish.

Findings

9 percentage point improvement on RAG Reader task

Strong performance in Polish MT-Bench, especially in Reasoning and Role-playing

Sets new benchmarks in Polish NLP tasks

Abstract

We introduce Bielik 7B v0.1, a 7-billion-parameter generative text model for Polish language processing. Trained on curated Polish corpora, this model addresses key challenges in language model development through innovative techniques. These include Weighted Instruction Cross-Entropy Loss, which balances the learning of different instruction types, and Adaptive Learning Rate, which dynamically adjusts the learning rate based on training progress. To evaluate performance, we created the Open PL LLM Leaderboard and Polish MT-Bench, novel frameworks assessing various NLP tasks and conversational abilities. Bielik 7B v0.1 demonstrates significant improvements, achieving a 9 percentage point increase in average score compared to Mistral-7B-v0.1 on the RAG Reader task. It also excels in the Polish MT-Bench, particularly in Reasoning (6.15/10) and Role-playing (7.83/10) categories. This model represents a substantial advancement in Polish language AI, offering a powerful tool for diverse linguistic applications and setting new benchmarks in the field.