Update README.md

README.md

@@ -22,17 +22,73 @@ This model is a fine-tuned version of [unsloth/meta-llama-3.1-8b-instruct-bnb-4b
 
 ## Model description
 
-More information needed
-
-## Intended uses & limitations
-
-More information needed
+This model was trained on Successful episodes of the top 3 model similar to [D20002](clembench-playpen/meta-llama-Meta-Llama-3.1-8B-Instruct_SFT_E1_D20002) but instead of using the whole episode as input,
+each episode was split into conversation pieces. 
+
+e.g. 
+```json
+[
+{
+  role: 'user'
+  content: '...'
+},
+{
+  role: 'assistant'
+  content: '...'
+},
+{
+  role: 'user'
+  content: '...'
+},
+{
+  role: 'assistant'
+  content: '...'
+},
+]
+```
+
+```json
+is split int:
+
+[
+{
+  role: 'user'
+  content: '...'
+},
+{
+  role: 'assistant'
+  content: '...'
+},
+```
+
+and
+
+```json
+[
+{
+  role: 'user'
+  content: '...'
+},
+{
+  role: 'assistant'
+  content: '...'
+},
+{
+  role: 'user'
+  content: '...'
+},
+{
+  role: 'assistant'
+  content: '...'
+},
+]
+```
 
 ## Training and evaluation data
 
-More information needed
+After splitting, the dataset contains about 4122 conversation bits accross all games.
 
-## Training procedure
+The Dataset ID is D30003
 
 ### Training hyperparameters