-
Notifications
You must be signed in to change notification settings - Fork 197
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update extract_score_and_reasons to work across providers #732
Conversation
I think the more structured messages should be preferable if available so I'm not sure about just using prompts always. |
Good point - updated to use prompt only for bedrock. |
import json | ||
|
||
body = json.dumps({"inputText": prompt}) | ||
body = json.dumps({"inputText": prompt, "textGenerationConfig": { | ||
"maxTokenCount": 4096, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Make a task to make these parameters configurable in a future PR?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@joshreini1 approved but see this note.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Added a TODO for this
* always use prompt instead of messages * use messages in base * use prompt in bedrock * move score to top of cot template, request entire template be used * remove dev * add TODO
* always use prompt instead of messages * use messages in base * use prompt in bedrock * move score to top of cot template, request entire template be used * remove dev * add TODO
Update structured extraction and COT prompting for better results across providers e.g. Bedrock, Ollama
Fixes: #716 and open email with amazon team, possibly: #609