Open qiuhaolu opened 9 months ago
I have also tried to use this model for testing with a medical subspecialty board exam. I have prev. results of having tested this exam with LLAMA-2-70b-chat and had OK results (much worse that chatgpt-3.5 or chatgpt-4 but much better than guessing.) I was hoping to see some improvements with this model - and hopefully close the gap to chatgpt-4 - as it seems to be trained for medical/biological questions. But I am was having trouble getting it to answer the multiple choice questions A/B/C/D with an answer. Instead it would go for long explanations. I have 300 question in my exam, I was hoping this model would be similar to LLAMA-2-70b-chat but better and my API driven test evaluation would have mostly worked. Alas, it is not the "chat" variant or for some reason not as helpful for my exam evaluation.
HI , have you tried fine tuning the model for this task .. I see in some replies they ask to fine the model for the task you want
HI , have you tried fine tuning the model for this task .. I see in some replies they ask to fine the model for the task you want
No, I haven't.
Hi, thank you for releasing the model. I believe it's a great resource for biomedical research and applications.
I am trying to prompt Meditron (7b, 70b-4b) for a medical NER task (using the example prompt, i.e., you are a helpful...) and couldn't get good results. Any suggestions on how to do NER with this model?