SPLICE Working Group: Large Language Models
Leaders: Juho Leinonen and Bita Akram.
Rationale: The emergence of LLMs and their forthcoming integration into CS education necessitates bringing together a community of CS education practitioners and researchers active in this area. This collaboration ensures LLMs' seamless and impactful incorporation into educational practices. The Splice LLM working group's primary objective is to foster a community where scholars can share LLM-related algorithms, datasets, reusable tools, and research findings to support CS instruction.
In particular, Splice LLM working group aims to support the CS education community through the following:
- Providing an infrastructure where LLM-based tools can be shared and reused.
- Providing a platform for sharing best practices for integrating LLMs for CS education support. This includes but is not limited to pedagogy design, prompt engineering, user studies, etc.
- Building an open dataset repository for 1) fine-tuning education-specific LLMs, 2) ew-shot prompting.
- Creating an open-source LLM along with an API to facilitate fine-tuning.
Get in Contact: Google Group