equivalence feedback gives in terms of user entered atoms, not expected atoms

Author: ashwin6-dev

evaluation_function/domain/evaluators.py

@@ -112,8 +112,9 @@ def evaluate_with_counterexample(self) -> tuple[bool, dict | None]:
                 v2 = FormulaEvaluator(self._formula2, a2).evaluate()
                 if v1 != v2:
                     if first_counterexample is None:
+                        # Use response formula's atom names (atoms1) so feedback speaks in the user's variables
                         first_counterexample = {
-                            "assignment": {atoms2[i].name: assignment_values[i] for i in range(n)},
+                            "assignment": {atoms1[j].name: assignment_values[perm[j]] for j in range(n)},
                             "response_value": v1,
                             "expected_value": v2,
                         }