For other nerds wanting some ways to | OpenAI | Page 1

faint vale Dec 28, 2022, 4:35 AM

#

Trajectories = What Chat Names Sessions. Policy = Triggerwords (edited)

#

Action Space: Training Space Objectives (I.E. Summarization/etc)

#

State is like the Session, while Observation is like your Input.

#

The Reward System!

#

The RL Problem (RL is the Model we are talking about in terms of Chat Bot/AI Training)

#

Values and Functions

gloomy coyote Dec 28, 2022, 4:51 AM

#

Is there any way to bookmark threads?

faint vale Dec 28, 2022, 4:54 AM

#

I can send you the text in a DM if you like

#

How GPT might use NLP to search your input.

gloomy coyote Dec 28, 2022, 4:57 AM

#

faint vale I can send you the text in a DM if you like

Sure

#

Send please

faint vale Dec 28, 2022, 5:06 AM

#

Was making it a little more accessable for everyone

#

📎 RL_Model_explained_via_a_Chat_Bot_Not_chatGPT.pdf

#

📎 RL_Model_explained_via_a_Chat_Bot_Not_chatGPT.txt

faint vale Dec 28, 2022, 5:57 AM

#

https://openai.com/blog/chatgpt/

OpenAI

ChatGPT: Optimizing Language Models for Dialogue

We’ve trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests. ChatGPT is a sibling model to InstructGPT, which is trained to follow an instruction in

#For other nerds wanting some ways to