diff --git a/src/tokenizer/tokenizers/DefaultTokenizer.ts b/src/tokenizer/tokenizers/DefaultTokenizer.ts
index 16f9be7..85a552a 100644
--- a/src/tokenizer/tokenizers/DefaultTokenizer.ts
+++ b/src/tokenizer/tokenizers/DefaultTokenizer.ts
@@ -16,7 +16,7 @@ export class DefaultTokenizer implements Tokenizer {
   }
 
   recursiveTokenize(content: string): { word: string; offset: number }[] {
-    const trimIndexes = Array.from(content.matchAll(TRIM_CHAR_PATTERN))
+    const trimIndexes = Array.from(content.matchAll(this.getTrimPattern()))
       .sort((a, b) => a.index! - b.index!)
       .map((x) => x.index!);
     return [
diff --git a/src/tokenizer/tokenizers/EnglishOnlyTokenizer.test.ts b/src/tokenizer/tokenizers/EnglishOnlyTokenizer.test.ts
new file mode 100644
index 0000000..1c06da4
--- /dev/null
+++ b/src/tokenizer/tokenizers/EnglishOnlyTokenizer.test.ts
@@ -0,0 +1,37 @@
+import { createTokenizer } from "../tokenizer";
+import { TokenizeStrategy } from "../TokenizeStrategy";
+
+describe.each`
+  content                 | raw      | expected
+  ${"aa bb cc"}           | ${false} | ${["aa", "bb", "cc"]}
+  ${"Edith旧市街"}        | ${false} | ${["Edith"]}
+  ${"Edith旧city"}        | ${false} | ${["Edith", "city"]}
+  ${"イーディスold city"} | ${false} | ${["old", "city"]}
+  ${"イーディスold市街"}  | ${false} | ${["old"]}
+  ${"イーディス旧市街"}   | ${false} | ${[]}
+`("tokenize", ({ content, raw, expected }) => {
+  test(`tokenize(${content}, ${raw}) = ${expected}`, () => {
+    expect(
+      createTokenizer(TokenizeStrategy.ENGLISH_ONLY).tokenize(content, raw)
+    ).toStrictEqual(expected);
+  });
+});
+
+// XXX: Hack implementation. There are some problems especially in recursiveTokenize.
+describe.each`
+  content           | expected
+  ${"aa bb cc"}     | ${[{ word: "aa bb cc", offset: 0 }, { word: "bb cc", offset: 3 }, { word: "cc", offset: 6 }]}
+  ${"aa:bb:cc"}     | ${[{ word: "aa:bb:cc", offset: 0 }, { word: "bb:cc", offset: 3 }, { word: "cc", offset: 6 }]}
+  ${"## @smi"}      | ${[{ word: "## @smi", offset: 0 }, { word: "@smi", offset: 3 }]}
+  ${"Edith旧市街"}  | ${[{ word: "Edith旧市街", offset: 0 }, { word: "市街", offset: 6 }, { word: "街", offset: 7 }, { word: "", offset: 8 }]}
+  ${"Edith旧city"}  | ${[{ word: "Edith旧city", offset: 0 }, { word: "city", offset: 6 }]}
+  ${"ヒナold city"} | ${[{ word: "ヒナold city", offset: 0 }, { word: "ナold city", offset: 1 }, { word: "old city", offset: 2 }, { word: "city", offset: 6 }]}
+  ${"ヒナold市街"}  | ${[{ word: "ヒナold市街", offset: 0 }, { word: "ナold市街", offset: 1 }, { word: "old市街", offset: 2 }, { word: "街", offset: 6 }, { word: "", offset: 7 }]}
+  ${"ヒナ旧市街"}   | ${[{ word: "ヒナ旧市街", offset: 0 }, { word: "ナ旧市街", offset: 1 }, { word: "旧市街", offset: 2 }, { word: "市街", offset: 3 }, { word: "街", offset: 4 }, { word: "", offset: 5 }]}
+`("recursiveTokenize", ({ content, expected }) => {
+  test(`recursiveTokenize(${content}) = ${expected}`, () => {
+    expect(
+      createTokenizer(TokenizeStrategy.ENGLISH_ONLY).recursiveTokenize(content)
+    ).toStrictEqual(expected);
+  });
+});
diff --git a/src/tokenizer/tokenizers/EnglishOnlyTokenizer.ts b/src/tokenizer/tokenizers/EnglishOnlyTokenizer.ts
index e4b5339..eeca693 100644
--- a/src/tokenizer/tokenizers/EnglishOnlyTokenizer.ts
+++ b/src/tokenizer/tokenizers/EnglishOnlyTokenizer.ts
@@ -1,6 +1,7 @@
 import { DefaultTokenizer } from "./DefaultTokenizer";
 
-export const ENGLISH_ONLY_TRIM_CHAR_PATTERN = /[^a-zA-Z0-9_-]/g;
+// XXX: Hack implementation. There are some problems especially in recursiveTokenize.
+export const ENGLISH_ONLY_TRIM_CHAR_PATTERN = /[^a-zA-Z0-9_\-#@]/g;
 export class EnglishOnlyTokenizer extends DefaultTokenizer {
   getTrimPattern(): RegExp {
     return ENGLISH_ONLY_TRIM_CHAR_PATTERN;