fix(graders): modify threshold value #59
Conversation
Summary of ChangesHello @weidankong, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request addresses an inconsistency in the OpenJudge grading system where common graders return scores on a scale of 1 to 5, but their default success threshold was set to 0.7, expecting a 0-1 scale. The changes update the threshold to operate within the 1-5 range, setting the new default to 3, ensuring logical alignment between grader outputs and their evaluation criteria. Highlights
🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request correctly updates the default threshold for several common graders (CorrectnessGrader, HallucinationGrader, HarmfulnessGrader, InstructionFollowingGrader) from 0.7 to 3, aligning with the new 1-5 scoring range. The changes are consistently applied across docstrings, examples, and method signatures.
I've added a few suggestions to improve type consistency for the threshold parameter.
A more significant concern is that some of the corresponding unit tests do not seem to be updated. For instance, in tests/graders/common/test_correctness.py and tests/graders/common/test_harmfulness.py, tests still use old, now invalid, threshold values like 0.7 and 0.8. It is crucial to update these tests to reflect the new logic and ensure the test suite remains valid and effective. Please ensure all related tests are updated as part of this change.
| self, | ||
| model: BaseChatModel | dict, | ||
| threshold: float = 0.7, | ||
| threshold: float = 3, |
There was a problem hiding this comment.
The type hint for threshold is float, but the default value is an integer 3. For consistency and to avoid potential type-related issues, it's better to use a float literal 3.0 as the default value.
| threshold: float = 3, | |
| threshold: float = 3.0, |
| self, | ||
| model: BaseChatModel | dict, | ||
| threshold: float = 0.7, | ||
| threshold: float = 3, |
There was a problem hiding this comment.
The type hint for threshold is float, but the default value is an integer 3. For consistency and to avoid potential type-related issues, it's better to use a float literal 3.0 as the default value.
| threshold: float = 3, | |
| threshold: float = 3.0, |
| self, | ||
| model: BaseChatModel | dict, | ||
| threshold: float = 0.7, | ||
| threshold: float = 3, |
There was a problem hiding this comment.
The type hint for threshold is float, but the default value is an integer 3. For consistency and to avoid potential type-related issues, it's better to use a float literal 3.0 as the default value.
| threshold: float = 3, | |
| threshold: float = 3.0, |
| self, | ||
| model: BaseChatModel | dict, | ||
| threshold: float = 0.7, | ||
| threshold: float = 3, |
There was a problem hiding this comment.
The type hint for threshold is float, but the default value is an integer 3. For consistency and to avoid potential type-related issues, it's better to use a float literal 3.0 as the default value.
| threshold: float = 3, | |
| threshold: float = 3.0, |
3 participants
OpenJudge Version
0.2.0
Description
common graders returning 1~5, however threshold defaults to 0.7
Checklist
Please check the following items before code is ready to be reviewed.
pre-commit run --all-filescommand