I will share the input we use for sentiment classification, as well as the key to connect to GPT-3 via slack channel.
You will use POSTMAN for this task. I will share all details with screenshots with you.
We want to reduce the size of the input. This can be done by removing the sentence and its expected result, and then seeing whether GPT-3 gives the same result with the removed sentence. If yes, then leave it removed. If no, place it back.
We want to see if there are any sentences that are not properly classified. This is done by placing the sentence at the end of the input, and verifying the output. If output is bad, we have to place it as a training example.
You can repeat reducing and inserting as many times you want (within a limit of 2000 calls). The main purpose is to optimise the calls and to increase the performance.
Please do not make more than 2000 calls. Let me know if you need more. Also, do not change the engine. We have to use ada !
Do not post your results here, share it via slack.
I will share the input we use for sentiment classification, as well as the key to connect to GPT-3 via slack channel. You will use POSTMAN for this task. I will share all details with screenshots with you.
We want to reduce the size of the input. This can be done by removing the sentence and its expected result, and then seeing whether GPT-3 gives the same result with the removed sentence. If yes, then leave it removed. If no, place it back.
We want to see if there are any sentences that are not properly classified. This is done by placing the sentence at the end of the input, and verifying the output. If output is bad, we have to place it as a training example.
You can repeat reducing and inserting as many times you want (within a limit of 2000 calls). The main purpose is to optimise the calls and to increase the performance.
Please do not make more than 2000 calls. Let me know if you need more. Also, do not change the engine. We have to use ada !
Do not post your results here, share it via slack.