Files
ragflow/agent/test/dsl_examples/retrieval_and_generate.json
Jin Hai 5e05f43c3d Update default prompt (#11984)
### What problem does this PR solve?

New default prompt:

```
You are an intelligent assistant. Your primary function is to answer questions based strictly on the provided knowledge base.

**Essential Rules:**
- Your answer must be derived **solely** from this knowledge base: `{knowledge}`.
- **When information is available**: Summarize the content to give a detailed answer.
- **When information is unavailable**: Your response must contain this exact sentence: "The answer you are looking for is not found in the knowledge base!"
- **Always consider** the entire conversation history.
```

Also fix some grammar errors.

### Type of change

- [x] Refactoring

Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2025-12-17 12:57:24 +08:00

61 lines
2.3 KiB
JSON

{
"components": {
"begin": {
"obj":{
"component_name": "Begin",
"params": {
"prologue": "Hi there!"
}
},
"downstream": ["retrieval:0"],
"upstream": []
},
"retrieval:0": {
"obj": {
"component_name": "Retrieval",
"params": {
"similarity_threshold": 0.2,
"keywords_similarity_weight": 0.3,
"top_n": 6,
"top_k": 1024,
"rerank_id": "",
"empty_response": "Nothing found in dataset",
"kb_ids": ["1a3d1d7afb0611ef9866047c16ec874f"]
}
},
"downstream": ["generate:0"],
"upstream": ["begin"]
},
"generate:0": {
"obj": {
"component_name": "LLM",
"params": {
"llm_id": "deepseek-chat",
"sys_prompt": "You are an intelligent assistant. Please summarize the content of the knowledge base to answer the question. Please list the data in the knowledge base and answer in detail. When all knowledge base content is irrelevant to the question, your answer must include the sentence \"The answer you are looking for is not found in the knowledge base!\" Answers need to consider chat history.\n Here is the knowledge base:\n {retrieval:0@formalized_content}\n Above is the knowledge base.",
"temperature": 0.2
}
},
"downstream": ["message:0"],
"upstream": ["retrieval:0"]
},
"message:0": {
"obj": {
"component_name": "Message",
"params": {
"content": ["{generate:0@content}"]
}
},
"downstream": [],
"upstream": ["generate:0"]
}
},
"history": [],
"path": [],
"retrieval": {"chunks": [], "doc_aggs": []},
"globals": {
"sys.query": "",
"sys.user_id": "",
"sys.conversation_turns": 0,
"sys.files": []
}
}