Are there any finetuned open-source models that the paper mentioned?๐ค
oushu1zhangxiangxuan1 opened this issue ยท 2 comments
oushu1zhangxiangxuan1 commented
Are there any finetuned open-source models that the paper mentioned?๐ค
BeachWang commented
Hi,
Thank you for your interest in our work. In this work, we have conducted insightful experiments on fine-tuned open-source models. However, certain challenges, such as the loss of in-context learning capacity, remain open questions. We are actively pursuing solutions in this area and will release our model once we have successfully addressed this issue.
vincent317 commented
Hello,
Thank you for the great work!
Quick question: When you mean the loss of in-context learning capacity for finetuned models, are these models finetuned on few-shot data as well or zero-shot data only?