Reinforcement Learning example

@dkundel-openai
:)

README.md

@@ -163,9 +163,17 @@ The gpt-oss models are excellent for:
 
 # Fine-tuning
 
-Both gpt-oss models can be fine-tuned for a variety of specialized use cases.
+Both gpt-oss models can be fine-tuned for a variety of specialized use-cases by using [transformers](https://github.com/huggingface/transformers) and [Unsloth](https://docs.unsloth.ai/new/gpt-oss-how-to-run-and-fine-tune).
 
-This smaller model `gpt-oss-20b` can be fine-tuned on consumer hardware, whereas the larger [`gpt-oss-120b`](https://huggingface.co/openai/gpt-oss-120b) can be fine-tuned on a single H100 node.
+This smaller model `gpt-oss-20b` can be fine-tuned on consumer hardware, whereas the larger [`gpt-oss-120b`](https://huggingface.co/openai/gpt-oss-120b) can be fine-tuned on a single H100 GPU.
+
+You can learn more about fine-tuning gpt-oss from [Hugging Face](https://cookbook.openai.com/articles/gpt-oss/fine-tune-transfomers) or [Unsloth’s guide](https://docs.unsloth.ai/new/gpt-oss-how-to-run-and-fine-tune#fine-tuning-gpt-oss-with-unsloth).
+
+## Reinforcement Fine-tuning
+
+You can also train `gpt-oss` with reinforcement learning (RL).
+
+[OpenAI’s notebook](https://github.com/openai/gpt-oss/blob/main/examples/reinforcement-fine-tuning.ipynb) shows how you can train `gpt-oss-20b` with RL to autonomously solve the 2048 game.
 
 # Citation