Clarify telecom data refueling rule in manual#105
Open
carol-xrl wants to merge 2 commits intosierra-research:mainfrom
Open
Clarify telecom data refueling rule in manual#105carol-xrl wants to merge 2 commits intosierra-research:mainfrom
carol-xrl wants to merge 2 commits intosierra-research:mainfrom
Conversation
[fix issue sierra-research#104: Reward Logic Flaw: Hard-Coded 2GB Refuel Requirement Causes Incorrect Evaluation] by replaceing "The maximum amount of data that can be refueled is 2GB" with a clearer rule: "Whenever refueling is considered, the amount of data to be refueled must be exactly 2GB."
[fix issue#104: Reward Logic Flaw: Hard-Coded 2GB Refuel Requirement Causes Incorrect Evaluation] by replaceing "The maximum amount of data that can be refueled is 2GB" with a clearer rule: "Whenever refueling is considered, the amount of data to be refueled must be exactly 2GB."
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
This PR solve the Reward Logic Flaw mentioned in
issue#104by updating the telecom manual by replacing the previous wording:with the more explicit rule:
Only this line in the manual is modified.
Why This Change Is Needed
The previous manual described a maximum limit, suggesting that the assistant may refuel any amount up to 2GB.
However, the reward function uses a strict assertion:
meaning the evaluation framework requires exactly 2GB whenever refueling occurs.
This mismatch causes reasonable assistant behavior (e.g., refueling less than 2GB) to be incorrectly judged as failure, resulting in an inaccurate and unfair evaluation score.