IDs-C2I-Enc is a bi-encoder retrieval model specifically fine-tuned for the IDs subset of the CodeConvo dataset. (IDs is short for Internet-Drafts)
You can use MTEB to load this model (
source code).
To run the IR evaluation task (
source code).
You can reproduce this with
this script (reproduced here for convenience).
1@article{bian2025automated,
2 title={Automated Insights Into GitHub Collaboration Dynamics},
3 author={Bian, Jie and Arefev, Nikolay and M{\"u}hlh{\"a}user, Max and Welzl, Michael},
4 journal={IEEE Access},
5 volume={13},
6 pages={85526--85542},
7 year={2025}
8}
