Epiplexity-optimized LIMA with fixed focus term extraction

README.md

@@ -0,0 +1,33 @@
+---
+dataset_info:
+  features:
+  - name: prompt
+    dtype: string
+  - name: completion
+    dtype: string
+  - name: source
+    dtype: string
+  - name: task_type
+    dtype: string
+  - name: response_format
+    dtype: string
+  - name: style
+    dtype: string
+  - name: focus
+    dtype: string
+  - name: prefix
+    dtype: string
+  - name: conversations_original
+    list: string
+  splits:
+  - name: train
+    num_bytes: 5885737
+    num_examples: 1030
+  download_size: 3336682
+  dataset_size: 5885737
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+---

data/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f914060d39ca8f88e4b3ae527d0d6abbc861d193428dcca082a0a62369b9ff5c
+size 3336682