Add support for naming the AI entity differently (other than "AI", which seems to be hardcoded)

[shobith]

This will allow the LLMs to hold on to previous commands.

[hwchase17]

sorry im not sure i see the connection... how does this relate to holding onto previous commands?

[shobith]

I was noticing a lack of "stickiness" to the prompt, I was wondering if it was because the memory annotated the bot's messages as "AI" (and the bot is named something else, say, something like "Jeeves").

Anyway, I was wondering if it'd be useful to have that bit customizable as well, or if I should go to a customized memory class for this purpose.

[hwchase17]

Ah i see. So you are using the name "Jeeves" in the prompt? like instead of using the default of

"""The following is a friendly conversation between a human and an AI. The AI is talkative and provides lots of specific details from its context. If the AI does not know the answer to a question, it truthfully says it does not know.
Current conversation:
{history}
Human: {input}
AI:"""

you used something like

"""The following is a friendly conversation between a human and Jeeves. Jeeves is talkative and provides lots of specific details from its context. If Jeeves does not know the answer to a question, it truthfully says it does not know.
Current conversation:
{history}
Human: {input}
Jeeves:"""

If that is the case, then yes i think this makes a ton of sense! Can probably expose some optional params like "llm_output_name" and "input_name" which default to "AI" and "Human" respectively (to maintain current functionality)

If this isnt what you meant though please let me know, I'd love to try to understand as best as possible

[shobith]

Yup, that's exactly what I meant. I was able to get around it by more prompting though, so this may not be super important.

[dosubot[bot]]

Hi, @shobith. I'm helping the LangChain team manage their backlog and I wanted to let you know that we are marking this issue as stale.

Based on my understanding, you requested the ability to name the AI entity something other than "AI" in order to retain previous commands. There was some discussion between you and "hwchase17" about the lack of "stickiness" to the prompt and the possibility of adding optional parameters to customize the name of the AI entity. However, it seems that the issue remains unresolved.

Could you please confirm if this issue is still relevant to the latest version of the LangChain repo? If you respond, the LangChain team will be notified to take a look. Otherwise, if there is no response within 7 days, the issue will be automatically closed. Thank you!