Instruction-Tuning-Papers
FreeA trend starts from `Natrural-Instruction` (ACL 2022), `FLAN` (ICLR 2022) and `T0` (ICLR 2022).
FreeFree tier
About Instruction-Tuning-Papers
A GitHub repository that curates a reading list of papers on instruction tuning, a technique for teaching language models to follow natural language instructions to improve multi-task learning and generalization to unseen tasks. The list includes seminal papers such as Natrural-Instruction (ACL 2022), FLAN (ICLR 2022), and T0 (ICLR 2022), along with many other works from 2021 onwards, each with direct links to the original papers.
Key Features
Curated collection of instruction-tuning papers from 2021 onwards
Direct links to original papers on arXiv or other sources
Covers key milestones and influential works in the field
Includes an explanation of instruction-tuning concept and its goals
Pros & Cons
Pros
- Comprehensive and well-organized list of important papers
- Open source and free to access on GitHub
- Provides direct links to papers for easy reading
- Useful for both newcomers and experienced researchers in NLP
Cons
- Not an interactive tool; it is a static reading list
- May not include the most recent papers if not regularly updated
- Only references papers, with no code or detailed analysis
Best For
Research and literature review on instruction tuningStaying updated on the latest NLP trends and techniquesAcademic reference for papers on language model fine-tuningLearning about the evolution of multi-task learning and zero-shot generalization
FAQ
What is instruction tuning?
Instruction tuning teaches language models to follow natural language instructions (including prompts, positive or negative examples, and constraints) to perform better on multi-task learning and generalize to unseen tasks.
Is this reading list free to access?
Yes, the repository is hosted publicly on GitHub and is free for anyone to view and use.