Skip to content

Fix self.sentiments refering to subtopic #223

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.

Sign up for GitHub

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

Merged
merged 2 commits into from
Apr 30, 2024
Merged

Fix self.sentiments refering to subtopic #223

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
8b963a6
Fix self.sentiments refering to subtopic
bkb2135 Apr 30, 2024
dc38535
Select only one option out of the completion
p-ferreira Apr 30, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 7 additions & 12 deletions prompting/rewards/ordinal.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,17 +12,11 @@ def name(self) -> str:
def __init__(self, **kwargs):
super().__init__()
#TODO: Expand to allow for more than 3 classes (Must also adjust dataset/review.py)
self.sentiments = [
"casual",
"basic",
"silly",
"random",
"thoughtful",
"serious",
"rushed",
self.sentiments = [
"positive",
"neutral",
"negative",
]
#NOTE: These sentimens are not the same as the sentiments defined in the dataset/review.py file. These are the subtopic


def reward(self, reference: str, completions: List[str]) -> BatchRewardOutput:
"""Compute difference scores given a completion and reference pair."""
Expand All @@ -34,7 +28,8 @@ def reward(self, reference: str, completions: List[str]) -> BatchRewardOutput:

# Check if exactly one answer can be found in the completion
if sum(option in completion for option in classes) == 1:
reward = abs(classes.index(reference) - classes.index(completion))
answer = [option for option in classes if option in completion][0]
reward = abs(classes.index(reference) - classes.index(answer))
else:
reward = 0
timings.append(time.time() - t0)
Expand All @@ -47,4 +42,4 @@ def reward(self, reference: str, completions: List[str]) -> BatchRewardOutput:
"type": "math",
},
)
return output
return output