#small language models

try —

7 papers match

cs.CL2026

Models for minimalist RAG: B1ade 335M Embedding and 1B Parameter Small Language Models

Shreyas Subramanian, Mecit Gungor, Vikram Elango

The paper presents B1ade, a resource‑efficient retrieval‑augmented generation system that combines a 335M parameter embedding model built by fusing five pretrained encoders with a…

#retrieval-augmented generation#embedding models#small language models#reinforcement learning
cs.CL2026

CoTu at EXACT 2026: Neuro-Symbolic Reasoning for Transparent Educational QA

Quoc-Khang Tran, Minh-Thien Nguyen, Phu-An Thai +3

The paper presents a neuro‑symbolic Program‑of‑Thought system that uses a 4B language model to generate symbolic programs (Z3 encodings for regulation queries and Python code for p…

#neuro-symbolic reasoning#educational question answering#program-of-thought#explainable ai
cs.CL2026

Exploring Post-Training Alignment of Small Language Models for Biomedical Data-to-Text Generation: A Case Study of Medication Leaflet

Xi Yang, Guodong Liu, Chuqin Li +10

The paper compares several post‑training alignment methods for small language models on the task of converting biomedical data into patient‑friendly medication leaflets, showing th…

#biomedical text generation#small language models#post-training alignment#preference optimization
cs.CL2026

Evaluating Health Misinformation in Low-Resource Languages: Integrating Small Language Models with a Culturally-Sensitive Responsible NLP Framework (Bangla as a Case Study)

Farnaz Farid, Raihan Alam, Al Al-Areqi +5

The paper proposes a culturally-sensitive AI framework for detecting health misinformation in low-resource languages, using Bangla as a case study and evaluating small language mod…

#health misinformation#low-resource languages#bangla#small language models
cs.SE2026

Form, Not Content? A Preregistered, Placebo-Controlled Evaluation of Learned Error-Conditioned Self-Repair Through Prompts and Weights in Frozen Small Code Models

Mehmet Iscan

The paper introduces a placebo‑controlled evaluation framework (PoPE) to test whether frozen small code language models can use error feedback to self‑repair generated code, and re…

#code generation#self-repair#placebo-controlled evaluation#prompt engineering
cs.AI2026

Improving Molecular Property Prediction in Small Language Models Using Graph-based Tools

Konstantinos Bougiatiotis, Dimitrios Kelesis, Georgios Paliouras

The paper proposes a Context‑Augmented Prompting framework that lets small language models query a graph neural network expert for structural hints and explanatory subgraphs, impro…

#small language models#molecular property prediction#graph neural networks#prompt engineering