@vivilyhere's a repro:

%scala

val tableName = "delta.`/Volumes/<YOUR_PATH_HERE>/delete_me_test_table1`"

spark.sql(
  "INSERT INTO IDENTIFIER(:table1)(id,name) VALUES(1,'alice');",
  Map("table1" -> tableName)
).show()
//above succeeds.

spark.sql(
  "INSERT INTO IDENTIFIER(:table1)(id,name) VALUES(1,:name);",
  Map("table1" -> tableName, "name" -> "bob")
).show()
//above fails with 14.3 compute, succeeds with 15.4 ML compute

Note: this is the SQL I used elsewhere to create the table:

CREATE TABLE delta.`/Volumes/<YOUR_PATH_HERE>/delete_me_test_table1`
SELECT
-1 as id,
'initial' as name;