The fine art of fine-tuning: A structured review of advanced LLM fine-tuning techniques
Transformer-based models have consistently demonstrated superior accuracy compared to various traditional models across a range of downstream tasks. However, due to their large nature, training or fine-tuning them for specific tasks has heavy computational and memory demands. This causes the creatio...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-06-01
|
| Series: | Natural Language Processing Journal |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2949719125000202 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850218326042607616 |
|---|---|
| author | Samar Pratap Alston Richard Aranha Divyanshu Kumar Gautam Malhotra Anantharaman Palacode Narayana Iyer Shylaja S.S. |
| author_facet | Samar Pratap Alston Richard Aranha Divyanshu Kumar Gautam Malhotra Anantharaman Palacode Narayana Iyer Shylaja S.S. |
| author_sort | Samar Pratap |
| collection | DOAJ |
| description | Transformer-based models have consistently demonstrated superior accuracy compared to various traditional models across a range of downstream tasks. However, due to their large nature, training or fine-tuning them for specific tasks has heavy computational and memory demands. This causes the creation of specialized transformer-based models to be almost impossible in the generally present constrained scenarios. To tackle this issue and to make these large models more accessible, a plethora of techniques have been developed. In this study, we will be reviewing the types of techniques developed, their impacts and benefits concerning performance and resource usage along with the latest developments in the domain. We have broadly categorized these techniques into six key areas: Changes in Training Method, Changes in Adapter, Quantization, Parameter Selection, Mixture of Experts, and Application based methods. We collated the results of various techniques on common benchmarks and also evaluated their performance on different datasets and base models. |
| format | Article |
| id | doaj-art-45370e07d4124e008f07468bde0c6840 |
| institution | OA Journals |
| issn | 2949-7191 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Natural Language Processing Journal |
| spelling | doaj-art-45370e07d4124e008f07468bde0c68402025-08-20T02:07:47ZengElsevierNatural Language Processing Journal2949-71912025-06-011110014410.1016/j.nlp.2025.100144The fine art of fine-tuning: A structured review of advanced LLM fine-tuning techniquesSamar Pratap0Alston Richard Aranha1Divyanshu Kumar2Gautam Malhotra3Anantharaman Palacode Narayana Iyer4Shylaja S.S.5Department of Computer Science, PES University, Bengaluru, 560085, Karnataka, IndiaDepartment of Computer Science, PES University, Bengaluru, 560085, Karnataka, IndiaDepartment of Computer Science, PES University, Bengaluru, 560085, Karnataka, IndiaDepartment of Computer Science, PES University, Bengaluru, 560085, Karnataka, India; Corresponding author.JNResearch, Bengaluru, 560085, Karnataka, IndiaDepartment of Computer Science, PES University, Bengaluru, 560085, Karnataka, IndiaTransformer-based models have consistently demonstrated superior accuracy compared to various traditional models across a range of downstream tasks. However, due to their large nature, training or fine-tuning them for specific tasks has heavy computational and memory demands. This causes the creation of specialized transformer-based models to be almost impossible in the generally present constrained scenarios. To tackle this issue and to make these large models more accessible, a plethora of techniques have been developed. In this study, we will be reviewing the types of techniques developed, their impacts and benefits concerning performance and resource usage along with the latest developments in the domain. We have broadly categorized these techniques into six key areas: Changes in Training Method, Changes in Adapter, Quantization, Parameter Selection, Mixture of Experts, and Application based methods. We collated the results of various techniques on common benchmarks and also evaluated their performance on different datasets and base models.http://www.sciencedirect.com/science/article/pii/S2949719125000202AdapterFFTLLMLoRAMoEPEFT |
| spellingShingle | Samar Pratap Alston Richard Aranha Divyanshu Kumar Gautam Malhotra Anantharaman Palacode Narayana Iyer Shylaja S.S. The fine art of fine-tuning: A structured review of advanced LLM fine-tuning techniques Natural Language Processing Journal Adapter FFT LLM LoRA MoE PEFT |
| title | The fine art of fine-tuning: A structured review of advanced LLM fine-tuning techniques |
| title_full | The fine art of fine-tuning: A structured review of advanced LLM fine-tuning techniques |
| title_fullStr | The fine art of fine-tuning: A structured review of advanced LLM fine-tuning techniques |
| title_full_unstemmed | The fine art of fine-tuning: A structured review of advanced LLM fine-tuning techniques |
| title_short | The fine art of fine-tuning: A structured review of advanced LLM fine-tuning techniques |
| title_sort | fine art of fine tuning a structured review of advanced llm fine tuning techniques |
| topic | Adapter FFT LLM LoRA MoE PEFT |
| url | http://www.sciencedirect.com/science/article/pii/S2949719125000202 |
| work_keys_str_mv | AT samarpratap thefineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques AT alstonrichardaranha thefineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques AT divyanshukumar thefineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques AT gautammalhotra thefineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques AT anantharamanpalacodenarayanaiyer thefineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques AT shylajass thefineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques AT samarpratap fineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques AT alstonrichardaranha fineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques AT divyanshukumar fineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques AT gautammalhotra fineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques AT anantharamanpalacodenarayanaiyer fineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques AT shylajass fineartoffinetuningastructuredreviewofadvancedllmfinetuningtechniques |