fix: Unify Prompts (#254)

* remove cot
* fix prompt template
* fix pool matrix
* spiral matrix fixed

reasoning_gym/arithmetic/leg_counting.py

@@ -56,16 +56,6 @@ ANIMALS = {
 
 QUESTION_TEMPLATE = """Your task is to count how many legs there are in total when given a list of animals.
 
-Example:
-- Input: How many legs are there in total if you have 1 duck, 2 deers, 1 spider, 3 cows?
-- Output: 30
-- Explanation:
-    - Ducks have 2 legs each, so 1 duck has 2 legs.
-    - Deers have 4 legs each, so 2 deers have 8 legs.
-    - Spiders have 8 legs each, so 1 spider has 8 legs.
-    - Cows have 4 legs each, so 3 cows have 12 legs.
-    - Therefore, the total number of legs is 2 + 8 + 8 + 12 = 30
-
 Now, how many legs are there in total if you have {animals}?
 """
 

reasoning_gym/arithmetic/number_format.py

@@ -8,12 +8,7 @@ from ..factory import ProceduralDataset, register_dataset
 
 QUESTION_TEMPLATE = """Your task is to pick the largest/smallest number out of several options.
 
-Example
-- Input: Pick the largest number of the following candidates: 857575.23 8.975554e+05 887,555.62
-- Output: 8.975554e+05
-- Explanation:
-    - Sorting the numbers written in various notations we get: 857575.23 < 887,555.62 < 8.975554e+05
-    - Therefore, the largest number is 8.975554e+05
+Your output should be only the number of interest.
 
 Now, pick the {size} number of the following candidates: {numbers}
 """

reasoning_gym/arithmetic/power_function.py

@@ -9,20 +9,6 @@ from ..factory import ProceduralDataset, register_dataset
 
 QUESTION_TEMPLATE = """Your task is to compute an exponentiation of a number.
 
-Example:
-- Input: Compute 2^3
-- Output: 8
-- Explanation:
-    - 2^3 = 2 * 2 * 2 = 8
-    - Therefore, the final answer is 8
-
-Example:
-- Input: Compute 412.5^3
-- Output: 70189453.125
-- Explanation:
-    - 412.5^3 = 412.5 * 412.5 * 412.5 = 70189453.125
-    - Therefore, the final answer is 70189453.125
-
 Compute {base}^{exponent}
 """