Update README.md

Author: sylee0520

README.md

@@ -3,7 +3,7 @@ This is the official github for "Volcano: Mitigating Multimodal Hallucination th
 Volcano employs a single LMM to generate initial responses, feedback, and revisions, as well as decisions to accept revisions. It follows a sequential procedure of an iterative critique-revision-decide loop. <br>
 - [Paper]() <br>
 - Model weights ([7B](https://huggingface.co/kaist-ai/volcano-7b), [13B](https://huggingface.co/kaist-ai/volcano-13b))
-- [Datasets](https://huggingface.co/datasets/kaist-ai/volcano-train)
+- [Training dataset](https://huggingface.co/datasets/kaist-ai/volcano-train)
 ## News
 \[Nov 14, 2023\] We released the first version of Volcano! Check out the paper, model and training dataset.
 ## Overview