Large language models (LLMs) have recently become the leading source of answers for users’ questions online. Despite their ability to offer eloquent answers, their accuracy and reliability can pose a significant challenge. This is especially true for sensitive domains such as biomedicine, where there is a higher need for factually correct answers. This paper introduces a biomedical retrieval-augmented generation (RAG) system designed to enhance the reliability of generated responses. The system is based on a fine-tuned LLM for the referenced question-answering, where retrieved relevant abstracts from PubMed are passed to LLM’s context as input through a prompt. Its output is an answer based on PubMed abstracts, where each statement is referenced accordingly, allowing the users to verify the answer. Our retrieval system achieves an absolute improvement of 23% compared to the PubMed search engine. Based on the manual evaluation on a small sample, our fine-tuned LLM component achieves comparable results to GPT-4 Turbo in referencing relevant abstracts. We make the dataset used to fine-tune the models and the fine-tuned models based on Mistral-7B-instruct-v0.1 and v0.2 publicly available.

In this paper we present first study of Sentiment Analysis (SA) of Serbian novels from the 1840-1920 period. The preparation of sentiment lexicon was based on three existing lexicons: NRC, AFFIN and Bing with additional extensive corrections. The first phase of dataset refinement included filtering the word that are not found in Serbian morphological dictionary and in second automatic POS tagging and lemma were manually corrected. The polarity lexicon was extracted and transformed into ontolex-lemon and published as initial version. The complex inflection system of Serbian language required expansion of sentiment lexicon with inflected forms from Serbian morphological dictionaries. Set of sentences for SA was extracted from 120 novels of Serbian part of ELTeC collection, labelled for polarity and used for several model training. Several approaches for SA are compared, starting with for variation of lexicon based and followed by Logistic Regression, Naive Bayes, Decision Tree, Random Forest, SVN and k-NN. The comparison with models trained on labelled movie reviews dataset indicates that it can not successfully be used for sentiment analysis of sentences in old novels.