small language model

PENGEMBANGAN KARTU UNO BIOLOGI (UBI) DENGAN CHALLANGE BOX DAN QR CODE PADA MATERI GENETIKA PEWARISAN SIFAT DI SMA

ABSTRACT The topic of genetic inheritance is conceptual and abstract in nature, requiring learning media that can connect scientific concepts with concrete and interactive learning activities. This study aimed to develop Biology UNO Cards (UBI) incorporating a Challenge Box and QR Code for teaching genetic inheritance...

Have scientists explained the emergence of languages by anatomical differences in the tongue and throat?

Did Scientists Really Explain the Origin of Languages through Anatomical Differences of the Tongue and Throat? Spoiler: No. This article, along with the study it refers to, is a textbook example of materialistic bankruptcy. They are attempting to explain the origin of an ultra-complex information system (which language...

Traditional Machine Learning versus Large Language Models for Predicting Ship Turnaround Times at Brazilian Ports

Predicting ship turnaround time at ports is an operationally valuable regression problem on heterogeneous, text-rich tabular data. We ask whether Large Language Models (LLMs) can match purpose-built Machine Learning (ML) on this task, using a leakage-controlled pipeline over seven years of Brazilian (ANTAQ) data. On a...

Den Übergang von Mündlichkeit zur Schriftlichkeit digital unterstützen

The transition from spoken to written language in German as a second language is a demanding but essential process and requires targeted support. Digital tools can support this transition; however, teachers require subject-specific digital competencies to effectively facilitate this process. The project DigKompDaZ_Fach...

A Benchmark of Small Language Models for Local Text-to-SQL over SQLite

Embeddings Are Overpriced? A Frozen 16 MB Vocabulary for Small and Ternary Language Models

Prompt Engineering for Small Language Models: Evaluating ICL for Portuguese Sentiment Analysis

The In-Context Learning (ICL) paradigm enables adapting LLMs without parameter tuning. This work evaluates the impact of prompt engineering on 9B models for Brazilian Portuguese sentiment classification, comparing six prompting formats (fewand zero-shot) across three linguistically diverse models (Gemma2-9B-it, Boto-9B...

Prevalent knowledge omissions: An unseen risk of LLMs in patient-friendly adaptation of medical literature

Large language models (LLMs) can produce patient-friendly adaptations, such as simplifying specialized medical literature for readers with a middle school education. However, any distortion of knowledge during this process may have serious consequences. Current evaluations overlook such risks. In this study, we propose...

Hallucination Detection Middleware using RAG Based Model

Retrieval-augmented generation (RAG) gives a language model access to external evidence, but retrieval alone does not ensure that a generated answer is supported by that evidence. This paper presents VeriRAG, a middleware-oriented design that separates query screening and evidence retrieval from answer verification. Ac...

Textual World Models for Agent Security: Modelling the Internal State of an LLM Agent from Text Alone

An AI agent chooses its own path to reach an objective, so its security and compliance depend on its internal state, not only on its inputs and outputs. We ask whether the activations of the agent's large language model (LLM), produced when it reads the context, are a sufficient observation of this state to build a mod...

Semantic Macro-Themes in Sertanejo Universitário Lyrics: Topic Modeling with BERTopic

Music lyrics provide a rich textual record of affective, social, and cultural themes, whose organization can be explored at scale through natural language processing. This paper investigates the internal thematic organization of 5,037 stanzas extracted from 959 Brazilian country songs released between 2005 and 2025 (se...

OpenCommons: A Distributed, Meaning-Indexed Concept Commons for Accelerating Scientific and Algorithmic Discovery in AI Agents

When swarms of autonomous AI agents solve complex coding and scientific discovery tasks, they reset their memory between sessions. As a result, agents waste compute repeatedly re-deriving the same mathematical algorithms from scratch while copying subtle numerical errors. Existing skill libraries either store unverifie...