-
-
Notifications
You must be signed in to change notification settings - Fork 642
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
Fixed up Max Mutual information example #842
Conversation
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.
❌ Changes requested. Reviewed everything up to e76e619 in 30 seconds
More details
- Looked at
270
lines of code in1
files - Skipped
1
files when reviewing. - Skipped posting
0
drafted comments based on config settings.
Workflow ID: wflow_0IXi8nKOi8ydGQyI
Want Ellipsis to fix these issues? Tag @ellipsis-dev
in a comment. You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet
mode, and more.
@@ -1,7 +1,260 @@ | |||
--- |
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.
The new markdown file max_mutual_information.md
should be added to mkdocs.yml
to ensure it is included in the site navigation.
Deploying instructor with Cloudflare Pages
|
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.
👍 Looks good to me! Incremental review on 5358961 in 38 seconds
More details
- Looked at
13
lines of code in1
files - Skipped
0
files when reviewing. - Skipped posting
1
drafted comments based on config settings.
1. mkdocs.yml:256
- Draft comment:
The navigation title formax_mutual_information.md
has been changed to 'Use Ensembles To Test Prompts'. If the document specifically focuses on the Max Mutual Information method, consider reverting this to a more accurate title, such as 'Max Mutual Information'.
- Max Mutual Information: 'prompting/ensembling/max_mutual_information.md'
- Reason this comment was not posted:
Confidence of 0% on close inspection, compared to threshold of 85%.
Workflow ID: wflow_1V9pckpXEzO3qqyt
You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet
mode, and more.
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.
hard to imagine when this will ever be used unforch...
I think unless you have the log probs, the implementation here is more of an example of LLM-as-a-judge to score different templates. I feel like this is just something to try but agree that few shot is going to give you much more mileage rather than messing around with prompts like this. |
Summary:
Updated
docs/prompting/ensembling/max_mutual_information.md
with detailed explanation, formulas, and Python implementation example for Max Mutual Information method.Key points:
docs/prompting/ensembling/max_mutual_information.md
with a detailed explanation of the Max Mutual Information method.prompt_template_1
andprompt_template_2
) for comparison.generate_response
,generate_responses
,calculate_entropy
, andcalculate_mutual_information
to compute and compare mutual information scores.Generated with ❤️ by ellipsis.dev