Paper: arXiv 2412.08179
Authors: Hai-Thien To, Tien-Cuong Bui, Van-Duc Le
Abstract
Financial analysis relies heavily on the interpretation of earnings reports to assess company performance and guide decision-making. Traditional methods for generating such analyzes require significant financial expertise and are often time-consuming. With the rapid advancement of Large Language Models (LLMs), domain-specific adaptations have emerged for financial tasks such as sentiment analysis and entity recognition. This paper introduces RAG-IT (Retrieval-Augmented Instruction Tuning), a novel framework designed to automate the generation of earnings report analysis through an LLM fine-tuned specifically for the financial domain. Our approach integrates retrieval augmentation with instruction-based fine-tuning to enhance factual accuracy, contextual relevance, and domain adaptability. We construct a sector-specific financial instruction dataset derived from semiconductor industry documents to guide the LLM adaptation to specialized financial reasoning. Using NVIDIA, AMD, and Broadcom as representative companies, our case study demonstrates that RAG-IT substantially improves a general-purpose open-source LLM and achieves performance comparable to commercial systems like GPT-3.5 on financial report generation tasks. This research highlights the potential of retrieval-augmented instruction tuning to streamline and elevate financial analysis automation, advancing the broader field of intelligent financial reporting.
Complexity vs Empirical Score
- Math Complexity: 2.0/10
- Empirical Rigor: 8.0/10
- Quadrant: Street Traders — empirical and applied, lower math complexity
Why this score: The paper uses minimal advanced mathematics, focusing on instruction tuning and retrieval augmented generation methodologies without heavy formulas or derivations. It demonstrates strong empirical rigor by constructing a sector-specific dataset, fine-tuning open-source LLMs, and comparing performance against commercial systems like GPT-3.5 on financial report generation tasks.
Research Flowchart
flowchart TD
A["Research Goal: Automate Financial Analysis<br>for Semiconductor Earnings Reports"] --> B
subgraph B ["Methodology: RAG-IT Framework"]
direction TB
B1["Data Collection:<br>Sector-Specific Documents"] --> B2["Instruction Tuning:<br>Fine-tune LLM on Financial Dataset"] --> B3["Retrieval Augmentation:<br>Integrate Knowledge Base during Inference"]
end
B --> C["Computational Process:<br>Generate Analysis for NVIDIA, AMD, Broadcom"]
C --> D["Key Findings:<br>1. Substantially improves open-source LLMs<br>2. Matches commercial systems (GPT-3.5)<br>3. Enhances factual accuracy & relevance"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes RAG-IT, a retrieval‑augmented instruction tuning framework for automating earnings report analysis with a financially fine‑tuned LLM.
Source sentence
This paper introduces RAG-IT (Retrieval-Augmented Instruction Tuning), a novel framework designed to automate the generation of earnings report analysis through an LLM fine-tuned specifically for the financial domain.
Assumes 1
- The earnings reports from Nvidia, AMD, and Broadcom in Q3 2023 are representative for evaluating the model. data
Source sentence
We train our model on the financial instruction dataset generated on earnings reports from two companies in the semiconductor sector, Nvidia and AMD, in the third quarter of 2023 and evaluate the earnings report of Broadcom Inc. in
Data
- financial instruction dataset text-news unknown
Source sentence
We train our model on the financial instruction dataset generated on earnings reports from two companies in the semiconductor sector, Nvidia and AMD, in the third quarter of 2023 and evaluate the earnings report of Broadcom Inc. in
Costs
Not applicable
Finds 2
- RAG-IT substantially improves a general‑purpose open‑source LLM and achieves performance comparable to GPT‑3.5 on financial report generation tasks. performance
Source sentence
our case study demonstrates that RAG-IT substantially improves a general‑purpose open‑source LLM and achieves performance comparable to commercial systems like GPT-3.5 on financial report generation tasks.
- The financially‑augmented LLM balances GPT‑3.5’s reasoning depth with Llama‑2’s fine‑tuning adaptability, integrating retrieval context for factual correctness and fluency. performance
Source sentence
Overall, our Financially‑Augmented LLM achieves a balance between GPT-3.5’s reasoning depth and Llama-2’s fine‑tuning adaptability. The model effectively integrates retrieval context to ensure factual correctness while maintaining natural linguistic fluency.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 13 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (18)
- Retrieval-augmented generation for knowledge-intensive nlp tasks (2020)
- Qlora: Efficient finetuning of quantized llms (2023)
- Llama 2: Open foundation and fine-tuned chat models (2023) — link
- Alpaca: A strong, replicable instruction-following model (2023)
- Visual instruction tuning (2023)
- BloombergGPT: A large language model for finance, 2023 (2024)
- FinGPT: Open-Source Financial Large Language Models — → 2306.06031
- Lora: Low-rank adaptation of large language models. (2022)
and 10 more
- Large Language Models in Equity Markets: Applications, Techniques, and Insights (2025)
- Leveraging Large Language Models in Finance: Pathways to Responsible Adoption (2025)
- Towards Alignment-Centric Paradigm: A Survey of Instruction Tuning in Large Language Models (2025) — link
- The Construction of Instruction-Tuned LLMs for Finance without Instruction Data Using Continual Pretraining and Model Merging — → 2409.19854
- Jafin: Japanese financial instruction dataset (2024)
- QuantMCP: Grounding Large Language Models in Verifiable Financial Reality (2025) — link
- Can Large language model analyze financial statements well? (2025)
- KPIs: What are key performance indicators? Types and examples (2024)
- 30 Financial Metrics and KPIs to Measure Success in 2025 (2025) — link
- KEY PERFORMANCE INDICATORS AND ANALYSTS’EARNINGS FORECAST ACCURACY: AN APPLICATION OF CONTENT ANALYSIS. (2011)
2 of 18 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.