RAG, CAG, and Long-Context Prompting with a 4-bit Quantized 7B Language Model for Question Answering on Higher-Education Regulations: Data, Code, and Results

Data, code, and analysis outputs for a comparison of retrieval-augmented generation (RAG), cache-augmented generation (CAG), and long-context prompting (LC) with 4-bit (NF4) quantized Qwen2.5-7B-Instruct on question answering over 17 Indonesian higher-education regulations (105 single-article questions; reference answers are verbatim quotations verified by script; no human raters or LLM judge). The archive contains the corpus, the QA set, raw outputs and per-item metrics of the main run and of a protocol-equalized rerun, analysis scripts, the system prompt, and environment versions. Greedy decoding; NVIDIA Tesla T4 (16 GB). This archive is not a preregistration.

Authors

Publication Details

Journal
Open Science Framework
Published
2026-11-12
DOI
https://doi.org/10.17605/osf.io/2qzc6
Primary Topic
Topic Modeling
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

RAG, CAG, and Long-Context Prompting with a 4-bit Quantized 7B Language Model for Question Answering on Higher-Education Regulations: Data, Code, and Results

Iwan Setiawan
Open Science Framework
Topic Modeling
article

RAG, CAG, and Long-Context Prompting with a 4-bit Quantized 7B Language Model for Question Answering on Higher-Education Regulations: Data, Code, and Results

Iwan Setiawan
article en

Abstract

Data, code, and analysis outputs for a comparison of retrieval-augmented generation (RAG), cache-augmented generation (CAG), and long-context prompting (LC) with 4-bit (NF4) quantized Qwen2.5-7B-Instruct on question answering over 17 Indonesian higher-education regulations (105 single-article questions; reference answers are verbatim quotations verified by script; no human raters or LLM judge). The archive contains the corpus, the QA set, raw outputs and per-item metrics of the main run and of a protocol-equalized rerun, analysis scripts, the system prompt, and environment versions. Greedy decoding; NVIDIA Tesla T4 (16 GB). This archive is not a preregistration.

Open Science Framework
Topic Modeling
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon