chore(weave): Public docs for MVP Guardrails #3344
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Preview this PR with FeatureBee: https://beta.wandb.ai/?betaVersion=928bbeaacea5530046df91344e5fd1200a2d13b4 |
andrewtruong
reviewed
Jan 8, 2025
| Everything is built around the concept of a `Scorer`. A `Scorer` is an instance of a sublcass of the `Scorer` class - particularly one that exposes a `score` method. See [Evaluation Metrics](./scorers.md) for more information. | ||
|
|
||
| ### Monitors | ||
| Now, many times you will want to apply a Scorer directly after calling an Op. This can be achieved by using the `apply_scorer` method on the `Call` object. |
andrewtruong
reviewed
Jan 8, 2025
There was a problem hiding this comment.
may want to strip the outputs of this nb
andrewtruong
approved these changes
Jan 8, 2025
tssweeney
commented
Jan 10, 2025
docs/notebooks/guardrails_example.py
Outdated
There was a problem hiding this comment.
Remove me - i am now a notebook
J2-D2-3PO
reviewed
Jan 10, 2025
|
|
||
|  | ||
|
|
||
| Weave provides a robust framework for implementing safety controls and monitoring systems in LLM applications through a unified scoring system. This guide explains how to leverage Scorers as both guardrails for active intervention and monitors for passive evaluation in production environments. |
There was a problem hiding this comment.
To match casing elsewhere (unless you meant the class)
Suggested change
| Weave provides a robust framework for implementing safety controls and monitoring systems in LLM applications through a unified scoring system. This guide explains how to leverage Scorers as both guardrails for active intervention and monitors for passive evaluation in production environments. | |
| Weave provides a robust framework for implementing safety controls and monitoring systems in LLM applications through a unified scoring system. This guide explains how to leverage scorers as both guardrails for active intervention and monitors for passive evaluation in production environments. |
|
|
||
| ## Monitors | ||
|
|
||
| While guardrails provide active intervention, monitors offer passive evaluation and tracking of LLM operations. They are crucial for long-term quality assurance and system improvement. |
There was a problem hiding this comment.
Question: Is the term "Guardrails" or "guardrails"? I've seen both floating around.
There was a problem hiding this comment.
i'll just keep it lower case for now
|
@tssweeney depending on timing, it might sense to merge this into #3066 and ship everything at once. With a little bit of modification, I think this PR would logically fit into the new structure. See #3066 |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Depends on: