Skip to content

Fix MCQ API answer-format prompt #152

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

Open
sylvesterkaczmarek wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Fix MCQ API answer-format prompt #152

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
20 changes: 19 additions & 1 deletion project/common/nanoeval/nanoeval/solvers/mcq_api.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,22 @@
""".strip()


def _answer_format_instruction(answer_keys: Sequence[str], allow_multiple_choices: bool) -> str:
answer_choices = ", ".join(answer_keys)
if allow_multiple_choices:
final_line_format = "'Answer: $LETTER1, $LETTER2, ...' (without quotes)"
choice_instruction = f"where each LETTER is one of {answer_choices}"
else:
final_line_format = "'Answer: $LETTER' (without quotes)"
choice_instruction = f"where LETTER is one of {answer_choices}"

return (
"Answer the following multiple choice question. The last line of your response should be "
f"of the following format: {final_line_format} {choice_instruction}. "
"Think step by step before answering."
)


@chz.chz
class OpenAIAPIMCQSolver(MCQSolver[Answer]):
model: str
Expand Down Expand Up @@ -94,7 +110,9 @@ async def solve(self, task: MCQTask) -> Answer:
{"role": "system", "content": "You are a helpful assistant."},
{
"role": "user",
"content": "Answer the following multiple choice question. The last line of your response should be of the following format: 'Answer: $LETTER' (without quotes) where LETTER is one of {answer_choices}. Think step by step before answering.",
"content": _answer_format_instruction(
list(letters_to_answers.keys()), question.allow_multiple_choices
),
},
{"role": "user", "content": prompt},
]
Expand Down