Refine Large Language Model Fine-tuning via Instruction Vector
Fine-tuning large language models (LLMs) can cause them to lose their general capabilities. However, the intrinsic mechanisms behind such forgetting remain unexplored. In this paper, we begin by examining this phenomenon by focusing on knowledge understanding and instruction following, with the latter identified as the main contributor to forgetting during fine-tuning. Consequently, we propose the Instruction Vector (IV) framework to capture model representations highly related to specific instruction-following capabilities, thereby making it possible to understand model-intrinsic forgetting. Through the analysis of IV dynamics pre and post-training, we suggest that fine-tuning mostly adds specialized reasoning patterns instead of erasing previous skills, which may appear as forgetting. Building on this insight, we develop IV-guided training, which aims to preserve original computation graph, thereby mitigating catastrophic forgetting. Empirical tests on three benchmarks confirm the efficacy of this new approach, supporting the relationship between IVs and forgetting. Our code will be made available soon.
Code (0)
등록된 구현이 없습니다.
Tasks
Instruction FollowingLanguage ModelingLanguage ModellingLarge Language ModelSimilar Papers 제목 키워드 기반
Self-Refine Instruction-Tuning for Aligning Reasoning in Language Models
The alignments of reasoning abilities between smaller and larger Language Models are largely conducted via Supervised Fine-Tuning (SFT) using demonstrations generated from robust Large Language Models (LLMs). Although th…
MathTowards Self-Refinement of Vision-Language Models with Triangular Consistency
Vision-Language Models (VLMs) integrate visual knowledge with the analytical capabilities of Large Language Models (LLMs) through supervised visual instruction tuning, using image-question-answer triplets. However, the p…
Mosaic-IT: Free Compositional Data Augmentation Improves Instruction Tuning
Finetuning large language models with a variety of instruction-response pairs has enhanced their capability to understand and follow instructions. Current instruction tuning primarily relies on teacher models or human in…
Data AugmentationDiversityInstruction FollowingSelective Reflection-Tuning: Student-Selected Data Recycling for LLM Instruction-Tuning
Instruction tuning is critical to large language models (LLMs) for achieving better instruction following and task adaptation capabilities but its success heavily relies on the training data quality. Many recent methods …
Data AugmentationInstruction FollowingTaCIE: Enhancing Instruction Comprehension in Large Language Models through Task-Centred Instruction Evolution
Large Language Models (LLMs) require precise alignment with complex instructions to optimize their performance in real-world applications. As the demand for refined instruction tuning data increases, traditional methods …
Diversity