罪を許す	2,824	27.1%	
罪を赦し	7,612	72.9%

赦す is not uncommon in certain contexts. I think notes are best approach.

@@ -12 +12 @@
-<ke_inf>&rK;</ke_inf>
+<ke_inf>&sK;</ke_inf>
@@ -16 +16 @@
-<ke_inf>&rK;</ke_inf>
+<ke_inf>&sK;</ke_inf>
@@ -26,0 +27 @@
+<s_inf>occ. written as 聴す</s_inf>
@@ -34,0 +36 @@
+<s_inf>also written as 赦す</s_inf>
@@ -42,0 +45 @@
+<s_inf>also written as 赦す</s_inf>

許す	1235723	97.2%
赦す	34950	2.7%
聴す	410	0.0%

The case for revisiting this one is not strong:
- in usage 許す dominates
- the JE dictionaries only have 許す for ゆるす. GG5 has it for all its 7 senses
- the major kokugos combine all three kanji forms in single entries. Daijirin adds tags such as 《許・赦》 and 《許・聴》 to some of the senses but 許 is used for all senses.
We could add notes such as "occ. 赦す" to some entries, but given the low level of usage of the other two forms, is it worth it?

@@ -11,0 +12 @@
+<ke_inf>&rK;</ke_inf>

Your kanji dictionary, supported  by the Unihan references
許 [Uni] 8a31  [音] キョ  [訓] ゆる.す もと  [英]  permit; approve More kanji details
赦 [Uni] 8d66  [音] シャ  [英]  pardon; forgiveness

EIJIRO is not 100% unambiguous but favours 赦す for "to pardon" and 許す for "to permit/tolerate"

I was told by a Japanese NS that there was a significant difference between 許す, which has a sense of 許可する and 赦す, which has a the sense of 恩赦する, and your kanji dictionary supports that assertion. I wonder if this would not warrant creating separate entries for 許す and 赦す...

Likewise, 聴す seems to be totally different from the above two... maybe this entry could be revisted?