| --- |
| language: |
| - en |
| license: apache-2.0 |
| pipeline_tag: text-generation |
| library_name: transformers |
| --- |
| |
| # SPARK: Adaptive Parametric Knowledge Participation for Faithful Retrieval-Augmented Generation |
|
|
|
|
| This repository contains the model checkpoints for **SPARK**, a lightweight framework designed to resolve knowledge conflicts in retrieval-augmented generation (RAG). |
|
|
| - **Paper:** [SPARK: Adaptive Parametric Knowledge Participation for Faithful Retrieval-Augmented Generation](https://huggingface.co/papers/2606.27786) |
| - **Repository:** [GitHub - OpenBMB/SPARK](https://github.com/OpenBMB/SPARK) |
|
|
| ## Method Overview |
|
|
| SPARK is a lightweight framework that learns adaptive knowledge participation policies for faithful retrieval-augmented generation. With the backbone LLM frozen, its participation policy dynamically regulates layer-wise FFN contributions, determining when and to what extent parametric knowledge participates during generation. Optimized through faithfulness-guided reinforcement learning using response-level rewards, SPARK mitigates catastrophic forgetting and better preserves general generation capabilities compared to full-parameter supervised fine-tuning. |
|
|
|
|
| ## Setup and Usage |
|
|
| Please refer to the official [GitHub Repository](https://github.com/OpenBMB/SPARK) for detailed environment setup, training, and evaluation scripts. |
|
|
| ## Citation |
|
|
| If you find this work useful, please cite the paper: |
|
|
| ```bibtex |
| @misc{li2026sparkadaptiveparametricknowledge, |
| title={SPARK: Adaptive Parametric Knowledge Participation for Faithful Retrieval-Augmented Generation}, |
| author={Ruochang Li and Pengcheng Huang and Zhenghao Liu and Yukun Yan and Huiyuan Xie and Yu Gu and Ge Yu and Maosong Sun}, |
| year={2026}, |
| eprint={2606.27786}, |
| archivePrefix={arXiv}, |
| primaryClass={cs.CL}, |
| url={https://arxiv.org/abs/2606.27786}, |
| } |
| ``` |
