DoG-Instruct: Towards Premium Instruction-Tuning Data via Text-Grounded Instruction Wrapping
By Yongrui Chen, Haiyun Jiang, Xinting Huang, Shuming Shi, Guilin Qi. 2023
16 citations recorded by the source corpus.
The improvement of LLMs’ instruction-following capabilities relies heavily on the availability of high-quality instruction-response pairs. Unfortunately, the current methods used to collect the pairs suffer from either unaffordable labor costs or severe hallucinations in the self-generation of LLM.To tackle these challenges, this paper proposes a scalable solution.It involves training LLMs to generate instruction-response pairs based on human-written documents, rather than relying solely on self-generation without context.Our proposed method not only exploits the advantages of human-written documents in reducing hallucinations but also utilizes an LLM to wrap the expression of documents, which enables us to bridge the gap between various document styles and the standard AI response.Experiments demonstrate that our method outperforms existing typical methods on multiple benchmarks.In particular, compared to the best-performing baseline, the LLM trained using our generated dataset exhibits a 10% relative improvement in performance on AlpacaEval, despite utilizing only 1/5 of its training data.Furthermore, a comprehensive manual evaluation validates the quality of the data we generated.
Subfields: Artificial Intelligence
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
- Xiao Wang
- Xianjun Yang
- L. Petzold
- William Yang Wang
- Xun Zhao
- Dahua Lin
- Qi Zhang
Citations in the corpus
- LongForm: Effective Instruction Tuning with Reverse Instructions (2023, neutral): “Thus, our dataset features a diverse collection of instructions (including human-written, template-based and LLM-generated), each paired with a human-written output.”
- CRAFT Your Dataset: Task-Specific Synthetic Dataset Generation Through Corpus Retrieval and Augmentation (2024, unknown): “Alternatively, some methods have started to use retrieval to enhance diversity: SynthesizRR (Divekar and Durrett, 2024) uses knowledge distillation from larger models while incorporating retrieved documents as in-context examples.”
- Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models (2023, background): “Large amounts of high-quality instruction-tuning data are expensive to collect, and scalable methods to automatically collect become popular (Chen et al., 2023; Zha et al., 2023; Zhao et al., 2023).”