diff --git a/docs/source/scrapers/llm.rst b/docs/source/scrapers/llm.rst index c22844d2..e76c56f1 100644 --- a/docs/source/scrapers/llm.rst +++ b/docs/source/scrapers/llm.rst @@ -132,10 +132,12 @@ We can also pass a model instance for the chat model and the embedding model. Fo azure_deployment="AZURE_OPENAI_EMBEDDINGS_DEPLOYMENT_NAME", openai_api_version="AZURE_OPENAI_API_VERSION", ) - + # Supposing model_tokens are 100K + model_tokens_count = 100000 graph_config = { "llm": { - "model_instance": llm_model_instance + "model_instance": llm_model_instance, + "model_tokens": model_tokens_count, }, "embeddings": { "model_instance": embedder_model_instance @@ -191,4 +193,4 @@ We can also pass a model instance for the chat model and the embedding model. Fo "embeddings": { "model_instance": embedder_model_instance } - } \ No newline at end of file + } diff --git a/scrapegraphai/prompts/search_internet_node_prompts.py b/scrapegraphai/prompts/search_internet_node_prompts.py index d786647b..f0508a53 100644 --- a/scrapegraphai/prompts/search_internet_node_prompts.py +++ b/scrapegraphai/prompts/search_internet_node_prompts.py @@ -11,4 +11,6 @@ You should return only the query string without any additional sentences. \n For example, if the user prompt is "What is the capital of France?", you should return "capital of France". \n If you return something else, you will get a really bad grade. \n +What you return should be sufficient to get the answer from the internet. \n +Don't just return a small part of the prompt, unless that is sufficient. \n USER PROMPT: {user_prompt}"""