open-thought/reasoning-gym
1
0
Fork 0
mirror of https://github.com/open-thought/reasoning-gym.git synced 2026-04-19 12:58:07 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Minor question template & score_answer improvements (#261)

Browse source 
* math prompt improvements
* ignore brackets in complex_arithmetic results
* improve additional instruction in prompt of polynomial_equations
* more strict tests for score_answer in polynomial_equations
* simplify special reward handling
* fix test_intermediate_integration
* fix sokoban dataset
* add common dataset score_answer consistency test
This commit is contained in:
Andreas Köpf 2025-03-04 21:55:09 +01:00 • committed by GitHub
parent 061282e373
commit 5d7fbac0ad
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
106 changed files with 403 additions and 507 deletions
Download patch file Download diff file Expand all files Collapse all files

5
reasoning_gym/arithmetic/number_format.py
View file

@ -65,14 +65,13 @@ class NumberFormatDataset(ProceduralDataset):
def score_answer(self, answer: Optional[str], entry: dict[str, Any]) -> float:
"""Overwrite this method in derived classes if a single oracle answer is not available."""
oracle_answer = entry["metadata"]["solution"]
if answer is not None and len(answer) > 0:
if isinstance(answer, str) and len(answer) > 0:
try:
answer = float(answer.strip().replace(",", ""))
if abs(answer - oracle_answer) < 1e-2:
return 1.0
return 0.01
except:
return 0.0
pass
return 0.0
def __getitem__(self, idx: int) -> dict: