From 1ef14a997e7a80eb114a2550e0dd4bb9fd45f6d5 Mon Sep 17 00:00:00 2001
From: ViperEkura <3081035982@qq.com>
Date: Mon, 21 Jul 2025 11:36:50 +0800
Subject: [PATCH] =?UTF-8?q?fix(chinese-instruct):=20=E4=BF=AE=E5=A4=8D=20L?=
 =?UTF-8?q?aTex=20=E5=85=AC=E5=BC=8F=E5=9C=A8=E6=95=B0=E6=8D=AE=E9=9B=86?=
 =?UTF-8?q?=E4=B8=AD=E7=9A=84=E6=98=BE=E7=A4=BA=E9=97=AE=E9=A2=98?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 chinese-instruct.py | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/chinese-instruct.py b/chinese-instruct.py
index 8543323..e8d551e 100644
--- a/chinese-instruct.py
+++ b/chinese-instruct.py
@@ -1,11 +1,20 @@
-# Mxode/Chinese-Reasoning-Distil-Data
-
 from datasets import DatasetDict
 from datasets import load_dataset, concatenate_datasets
 from utils import process_dataset
 
 
-def build_prompt(query, response) -> str:
+def build_prompt(query:str, response:str) -> str:
+    replacements = {
+        "\\[": "$$",
+        "\\]": "$$",
+        "\\(": "$",
+        "\\)": "$"
+    }
+
+    for old, new in replacements.items():
+        query = query.replace(old, new)
+        response = response.replace(old, new)
+         
     return f"<|user|> {query} <|system|> <bos>{response}<eos>\n"