paper-with-me

홈 › Papers

RECOST: External Knowledge Guided Data-efficient Instruction Tuning

2024-02-27 · Qi Zhang, Yiming Zhang, Haobo Wang, Junbo Zhao

In the current landscape of large language models (LLMs), the process of instruction tuning serves as an essential step. Considering the high computing power overhead, data-efficient instruction tuning was proposed to reduce the training data size in this process, aiming at selecting high-quality instructional data. Nevertheless, we argue that most current data-efficient instruction-tuning methods are highly dependent on the quality of the original instruction-tuning dataset. When it comes to datasets synthesized by LLMs, a common scenario in this field, dirty samples will even be selected with a higher probability than other samples. To address these challenges, we utilized external knowledge (relevant examples or paragraphs) to evaluate those samples synthesized by LLMs with an in-context-based relative predictive entropy. Based on the new metric, we proposed a framework, dubbed as \textbf{RECOST}, which integrates external-knowledge-base re-ranking and diversity-consistent sampling into a single pipeline. Through extensive experiments on several synthetic datasets (Alpaca and Alpaca-gpt4), we demonstrate the effectiveness of our method and achieve even better results with only \textbf{1\%} of the full dataset.

📄 PDF Abstract BibTeX arXiv:2402.17355

Code (0)

등록된 구현이 없습니다.

Tasks

DiversityRe-Ranking

Similar Papers 제목 키워드 기반

Landmark-Guided Knowledge for Vision-and-Language Navigation

2025-09-30 · Dongsheng Yang, Meiling Zhu, Yinfeng Yu arxiv

Vision-and-language navigation is one of the core tasks in embodied intelligence, requiring an agent to autonomously navigate in an unfamiliar environment based on natural language instructions. However, existing methods…

IKnow: Instruction-Knowledge-Aware Continual Pretraining for Effective Domain Adaptation

2025-10-23 · Tianyi Zhang, Florian Mai, Lucie Flek arxiv

Continual pretraining promises to adapt large language models (LLMs) to new domains using only unlabeled test-time data, but naively applying standard self-supervised objectives to instruction-tuned models is known to de…

Continual PretrainingDomain Adaptation

Effective Instruction Parsing Plugin for Complex Logical Query Answering on Knowledge Graphs

2024-10-27 · Xingrui Zhuo, Jiapu Wang, Gongqing Wu, Shirui Pan 외

Knowledge Graph Query Embedding (KGQE) aims to embed First-Order Logic (FOL) queries in a low-dimensional KG space for complex reasoning over incomplete KGs. To enhance the generalization of KGQE models, recent studies i…

Entity AlignmentKnowledge Graphs

Early Timestep Zero-Shot Candidate Selection for Instruction-Guided Image Editing

2025-04-18 · Joowon Kim, Ziseok Lee, Donghyeon Cho, Sanghyun Jo 외

Despite recent advances in diffusion models, achieving reliable image generation and editing remains challenging due to the inherent diversity induced by stochastic noise in the sampling process. Instruction-guided image…

Image Generation

FactLLaMA: Optimizing Instruction-Following Language Models with External Knowledge for Automated Fact-Checking

2023-09-01 · Tsun-Hin Cheung, Kin-Man Lam

Automatic fact-checking plays a crucial role in combating the spread of misinformation. Large Language Models (LLMs) and Instruction-Following variants, such as InstructGPT and Alpaca, have shown remarkable performance i…

Fact CheckingInstruction FollowingLanguage ModelingLanguage Modelling+1