3 ms·
I wonder if using prompts to get the sentiment in LLM is enough? So we do not need to do any fine-tuning anymore?
by chazeon 2y ago
I wonder if using prompts to get the sentiment in LLM is enough? So we do not need to do any fine-tuning anymore?
- t-writescode 2y agoI think you raise a reasonable question. I also think it *could* be less of a problem than you might think. If we treat the scale as arbitrary (which I think is a safe thing to do), then movement along the scale could be sufficient to ascertain *something*