diff --git a/reasoning_gym/games/knight_swap.py b/reasoning_gym/games/knight_swap.py index 08c23080..4b444be1 100644 --- a/reasoning_gym/games/knight_swap.py +++ b/reasoning_gym/games/knight_swap.py @@ -51,8 +51,8 @@ class KnightSwapConfig: max_nodes: int = 9 # Maximum number of squares on the board min_pieces: int = 2 # Minimum number of pieces per color max_pieces: int = 2 # Maximum number of pieces per color - min_steps: int = 3 # Minimum solution length - max_steps: int = 5 # Maximum solution length + min_steps: int = 4 # Minimum solution length + max_steps: int = 20 # Maximum solution length max_attempts: int = 50 # Maximum attempts for board generation and puzzle creation seed: Optional[int] = None size: int = 5 # Virtual dataset size diff --git a/reasoning_gym/games/puzzle24.py b/reasoning_gym/games/puzzle24.py index 8cb84019..a0b86d41 100644 --- a/reasoning_gym/games/puzzle24.py +++ b/reasoning_gym/games/puzzle24.py @@ -116,10 +116,8 @@ class Puzzle24Dataset(ProceduralDataset): try: answer = answer.strip() user_answer = int(parse_expr(answer)) - print(f"User answer: {user_answer}") solved = user_answer == 24 used_numbers = [int(num) for num in re.findall(r"\b\d+\b", answer)] - print(f"Used numbers: {used_numbers}") if len(used_numbers) != 4: reward = 0.01 elif any(num > self.config.max_value or num < self.config.min_value for num in used_numbers):