Update README.md

README.md

@@ -183,10 +183,11 @@ We use Best-of-64 as our evaluation metric. The weighting methods are different
 
 We evaluate **EurusPRM-Stage 1** and **EurusPRM-Stage 2** on **ProcessBench**. 
 The threshold is obtained by converting the original score of each step using sigmoid function and iterating to find the highest F1 on GSM8k sub-benchmark. The threshold for **EurusPRM-Stage 1** and **EurusPRM-Stage 2** is 0.5015 and 0.5005 respectively.
-For leveraging the capibility of **EurusPRM** better, we add ``Step K`` (where K is the actual index of the step) in front of each step in **ProcessBench**.
+
+To leverage the capibility of **EurusPRM** better, we add ``Step K`` (where K is the actual index of the step) in front of each step in **ProcessBench**.
 
 | Reward Model | GSM8k | MATH | OlympiadBench | Omni-Math | Avg |
-| --- | --- | --- | --- | --- | --- | --- |
+| --- | --- | --- | --- | --- | --- |
 | Math-Shepherd-PRM-7B | 47.9 | 29.5 | 24.8 | 23.8 | 31.5 |
 | RLHFlow-PRM-Mistral-8B | 50.4 | 33.4 | 13.8 | 15.8 | 28.4 |
 | RLHFlow-PRM-Deepseek-8B | 38.8 | 33.8 | 16.9 | 16.9 | 26.6 |