Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Balancing Continuous Pre-Training and Instruction Fine-Tuning: Optimizing Instruction-Following in LLMs

Ishan Jindal, Chandana Badrinath, Pranjal Bharti, Lakkidi Vinay, Sachin Dev Sharma

TL;DR

This study dives into the intricate relationship between continuous pre-training and instruction fine-tuning of the LLMs and investigates the impact of continuous pre-training on the instruction following abilities of both the base and its instruction finetuned model.

Abstract

Large Language Models (LLMs) for public use require continuous pre-training to remain up-to-date with the latest data. The models also need to be fine-tuned with specific instructions to maintain their ability to follow instructions accurately. Typically, LLMs are released in two versions: the Base LLM, pre-trained on diverse data, and the instruction-refined LLM, additionally trained with specific instructions for better instruction following. The question arises as to which model should undergo continuous pre-training to maintain its instruction-following abilities while also staying current with the latest data. In this study, we delve into the intricate relationship between continuous pre-training and instruction fine-tuning of the LLMs and investigate the impact of continuous pre-training on the instruction following abilities of both the base and its instruction finetuned model. Further, the instruction fine-tuning process is computationally intense and requires a substantial number of hand-annotated examples for the model to learn effectively. This study aims to find the most compute-efficient strategy to gain up-to-date knowledge and instruction-following capabilities without requiring any instruction data and fine-tuning. We empirically prove our findings on the LLaMa 3, 3.1 and Qwen 2, 2.5 family of base and instruction models, providing a comprehensive exploration of our hypotheses across varying sizes of pre-training data corpus and different LLMs settings.

Balancing Continuous Pre-Training and Instruction Fine-Tuning: Optimizing Instruction-Following in LLMs

TL;DR

This study dives into the intricate relationship between continuous pre-training and instruction fine-tuning of the LLMs and investigates the impact of continuous pre-training on the instruction following abilities of both the base and its instruction finetuned model.

Abstract

Large Language Models (LLMs) for public use require continuous pre-training to remain up-to-date with the latest data. The models also need to be fine-tuned with specific instructions to maintain their ability to follow instructions accurately. Typically, LLMs are released in two versions: the Base LLM, pre-trained on diverse data, and the instruction-refined LLM, additionally trained with specific instructions for better instruction following. The question arises as to which model should undergo continuous pre-training to maintain its instruction-following abilities while also staying current with the latest data. In this study, we delve into the intricate relationship between continuous pre-training and instruction fine-tuning of the LLMs and investigate the impact of continuous pre-training on the instruction following abilities of both the base and its instruction finetuned model. Further, the instruction fine-tuning process is computationally intense and requires a substantial number of hand-annotated examples for the model to learn effectively. This study aims to find the most compute-efficient strategy to gain up-to-date knowledge and instruction-following capabilities without requiring any instruction data and fine-tuning. We empirically prove our findings on the LLaMa 3, 3.1 and Qwen 2, 2.5 family of base and instruction models, providing a comprehensive exploration of our hypotheses across varying sizes of pre-training data corpus and different LLMs settings.

Paper Structure

This paper contains 24 sections, 5 equations, 2 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Resulted LLM
  4. Setting 1: Continuous Pre-training of Instruction-Tuned LLM
  5. Setting 2: Continuous Pre-training of Base LLM followed by Instruction Fine-tuning
  6. Instruction Residuals
  7. Experiments
  8. Datasets
  9. Pre-traininig Dataset
  10. Evaluation Dataset
  11. Language Model Architectures
  12. Results and Analysis
  13. Impact of Continual Pretraining
  14. Restore Instruction Capabilities
  15. Instruction Portability across LLM Families
  16. ...and 9 more sections

Figures (2)

  • Figure 1: Impact of continual pre-training on instruction following capability of LLaMa base and instruct models compared against residual technique.
  • Figure 2: Comparison of MMLU performance and compute costs across different models. Our model (marked in green) balances compute efficiency while maintaining competitive performance.