fix: remove duplicate safe_exp function, use exp_safe instead

Resolves #2149

The codebase had two similar exponentiation functions:
- exp_safe (line 172): Robust implementation with input clamping [-20, 20]
  and smart error handling (returns max_value for positive overflow)
- safe_exp (line 1597): Simple wrapper returning 0 on any error

This commit:
- Removes the duplicate safe_exp function from math.rs
- Updates run_epoch.rs to use exp_safe instead

exp_safe is the better choice as it:
1. Clamps input to prevent overflow in the first place
2. Returns max_value (not 0) when large positive inputs overflow
3. Has existing test coverage in tests/math.rs

Author: MkDev11

pallets/subtensor/src/epoch/math.rs

@@ -1592,8 +1592,3 @@ pub fn mat_ema_alpha(
 pub fn safe_ln(value: I32F32) -> I32F32 {
     ln(value).unwrap_or(I32F32::saturating_from_num(0.0))
 }
-
-/// Safe exp function, returns 0 if value is 0.
-pub fn safe_exp(value: I32F32) -> I32F32 {
-    exp(value).unwrap_or(I32F32::saturating_from_num(0.0))
-}

pallets/subtensor/src/epoch/run_epoch.rs

@@ -1536,7 +1536,7 @@ impl<T: Config> Pallet<T> {
 
         // sigmoid = 1. / (1. + e^(-steepness * (combined_diff - 0.5)))
         let sigmoid = one.saturating_div(
-            one.saturating_add(safe_exp(
+            one.saturating_add(exp_safe(
                 alpha_sigmoid_steepness
                     .saturating_div(I32F32::from_num(-100))
                     .saturating_mul(combined_diff.saturating_sub(I32F32::from_num(0.5))),