Misleading logs

[954-Ivory]

Checked other resources

• [X] I added a very descriptive title to this issue.
• [X] I searched the LangChain documentation with the integrated search.
• [X] I used the GitHub search to find a similar question and didn't find it.
• [X] I am sure that this is a bug in LangChain rather than my code.
• [X] The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).

Example Code

import os

from dotenv import load_dotenv
from langchain_core.globals import set_debug
from langchain_core.messages import HumanMessage, SystemMessage
from langchain_openai import ChatOpenAI

set_debug(True)
load_dotenv()
model = ChatOpenAI(
    api_key=os.getenv('OPENAI_API_KEY'),
    base_url=os.getenv('OPENAI_BASE_URL'),
    model="gpt-3.5-turbo"
)

messages = [
    SystemMessage(content="Translate the following from English into Italian"),
    HumanMessage(content="hi!"),
]

if __name__ == "__main__":
    print(model.invoke(messages))

Error Message and Stack Trace (if applicable)

[llm/start] [llm:ChatOpenAI] Entering LLM run with input:
{
  "prompts": [
    "System: Translate the following from English into Italian\nHuman: hi!"
  ]
}

Description

Shouldn't it be like this?

[llm/start] [llm:ChatOpenAI] Entering LLM run with input:
{
  "prompts": [
    "System": "Translate the following from English into Italian",
    "Human": "hi!"
  ]
}

System Info

I have tried it in two conda envs:

---

langchain 0.2.5 windows 11 python Python 3.11.9

---

langchain 0.1.10 windows 11 python Python 3.11.7

[wenngong]

The terminal output u see is just intermediate logs by ***CallBackHandler，and the handler function flatten the chat prompt messages into single string for logging.

so, it's not a bug I believe.

[954-Ivory]

The terminal output u see is just intermediate logs by ***CallBackHandler，and the handler function flatten the chat prompt messages into single string for logging.

so, it's not a bug I believe.

Yes, I have checked the network request and there is no problem. But such output will cause misunderstandings.