关于模型一般能力的获取疑问。

HITsz-TMG / Ext-Sub

Official implementation of our paper "Separate the Wheat from the Chaff: Model Deficiency Unlearning via Parameter-Efficient Module Operation". A model merge method for deficiency unlearning, compitable with huggingface peft (LoRA).

https://arxiv.org/abs/2308.08090

MIT License

3 stars 1 forks source link

关于模型一般能力的获取疑问。 #3

Closed fxb392 closed 4 days ago

fxb392 commented 2 weeks ago

为什么可以“通用能力v◦i的方向向量可以由ˆvi+和ˆvi−的单位向量相加得到”呢？如何解释呢？