Letting the weight of the model at 1 can work if you don't have hands on the images, else it may give some strange results. Thats why I suggest you set the weigth at 0.9 or 0.8, and if it still give bad hands, try to describe the position of the hands in the prompts (clenched hands, v sign, ...)
If you want to add a loras or embeddings in the prompt, try to lower their weight (especially for embeddings)
