Saved in:
Bibliographic Details
Main Authors: Situngkir, Hokky, Siringoringo, Kevin, Lumbantobing, Andhika Bernard
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.10006
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908877534003200
author Situngkir, Hokky
Siringoringo, Kevin
Lumbantobing, Andhika Bernard
author_facet Situngkir, Hokky
Siringoringo, Kevin
Lumbantobing, Andhika Bernard
contents This study presents TOBA-LM, a trilingual language model based on GPT-2 architecture with 1.2 billion parameters, trained on a corpus encompassing Indonesian, Batak, and Minangkabau using syllabic-agglutinative tokenization. The architecture integrates an Engram Memory mechanism, an adaptive n-gram-based memory system with a 500,000 x 768 embedding table that captures morphological dependencies through bigram and trigram pathways. Empirical results demonstrate a training efficiency of 80%, with the loss value dropping from 6.4 to 1.7996 in only 12,973 steps -- significantly faster than the conventional transformer architecture, which required over 70,000 steps to achieve comparable convergence. These findings confirm that the integration of external statistical memory substantially reduces computational requirements for developing regional language models under limited resources.
format Preprint
id arxiv_https___arxiv_org_abs_2603_10006
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Adaptive Engram Memory System for Indonesian Language Model: Generative AI Based on TOBA LM for Batak and Minang Language
Situngkir, Hokky
Siringoringo, Kevin
Lumbantobing, Andhika Bernard
Computation and Language
Computers and Society
68T50
I.2.7; E.4; F.2.2
This study presents TOBA-LM, a trilingual language model based on GPT-2 architecture with 1.2 billion parameters, trained on a corpus encompassing Indonesian, Batak, and Minangkabau using syllabic-agglutinative tokenization. The architecture integrates an Engram Memory mechanism, an adaptive n-gram-based memory system with a 500,000 x 768 embedding table that captures morphological dependencies through bigram and trigram pathways. Empirical results demonstrate a training efficiency of 80%, with the loss value dropping from 6.4 to 1.7996 in only 12,973 steps -- significantly faster than the conventional transformer architecture, which required over 70,000 steps to achieve comparable convergence. These findings confirm that the integration of external statistical memory substantially reduces computational requirements for developing regional language models under limited resources.
title Adaptive Engram Memory System for Indonesian Language Model: Generative AI Based on TOBA LM for Batak and Minang Language
topic Computation and Language
Computers and Society
68T50
I.2.7; E.4; F.2.2
url https://arxiv.org/abs/2603.10006