DSPy: The framework for programming—not prompting—language models
Python
update docs
Omar Khattab1b4bd6422c92
diff --git a/docs/docs/tutorials/math/index.ipynb b/docs/docs/tutorials/math/index.ipynbindex 86a0849b7..3e6c55894 100644--- a/docs/docs/tutorials/math/index.ipynb+++ b/docs/docs/tutorials/math/index.ipynb@@ -317,7 +317,7 @@"cell_type": "markdown","metadata": {},"source": [- "And lastly let's optimize our module. Since we want strong reasoning, we'll use the large GPT-4o as the teacher model (used to bootstrap reasoning for the small LM at optimization time) and as the proposer model (used to craft instructions).\n",+ "And lastly let's optimize our module. Since we want strong reasoning, we'll use the large GPT-4o as the teacher model (used to bootstrap reasoning for the small LM at optimization time) but not as the prompt model (used to craft instructions) or the task model (trained).\n","\n","GPT-4o will be invoked only a small number of times. The model involved directly in optimization and in the resulting (optimized) program will be GPT-4o-mini.\n","\n",