openai
About the 'logprobs' in the response object
Hi,
I find that the simulator will postprocess the response of 'text-davinci-003' using the response field 'logprobs'.
However, as I read the document of openai, the field 'logprobs' is going to be deprecated since the completion response object will be replaced by chat completion object, and also the model 'text-davince-003' is going to be deprecated.
I now have the access to the gpt-4 and gpt-3.5-turbo with chat completion response object. Is there any way to conduct the neuron-explainer using the two models? i.e. without the field 'logprobs'.
Or, is it necessary to call 'text-davince-003' and other models with completion response object that return 'logprobs'.
Thanks a lot!
Explanation will work without logpobs but our current scoring method won't, we will update if we have a better scoring method
On Wed, Nov 1, 2023 at 8:56 PM Youcheng Huang @.***> wrote:
Hi,
I find that the simulator will postprocess the response of 'text-davinci-003' using the response field 'logprobs'.
However, as I read the document of openai, the field 'logprobs' is going to be deprecated since the completion response object will be replaced by chat completion object, and also the model 'text-davince-003' is going to be deprecated.
I now have the access to the gpt-4 and gpt-3.5-turbo with chat completion response object. Is there any way to conduct the neuron-explainer using the two models? i.e. without the field 'logprobs'.
Or, is it necessary to call 'text-davince-003' and other models with completion response object that return 'logprobs'.
Thanks a lot!
— Reply to this email directly, view it on GitHub https://github.com/openai/automated-interpretability/issues/24, or unsubscribe https://github.com/notifications/unsubscribe-auth/ASWBKK7HMDWB3WNQBFDMQWDYCMKWPAVCNFSM6AAAAAA62HZWASVHI2DSMVQWIX3LMV43ASLTON2WKOZRHE3TGNBVHAYDANI . You are receiving this because you are subscribed to this thread.Message ID: @.***>
Hi @williamrs-openai , could you explain how the variables in L209 to L212 correspond to the logprobs results given in gpt4-1106? Especially choice["logprobs"]["text_offset"] made me confused.
Here I set logprobs=True and top_logprobs=1. choice.logprobs.content include:
By the way, how many should I set top_logprobs to?
