Initial upload: Japanese HH-RLHF → Anthropic DPO format (fixed line breaks)

README.md

@@ -0,0 +1,26 @@
+---
+dataset_info:
+  features:
+  - name: chosen
+    dtype: string
+  - name: rejected
+    dtype: string
+  - name: prompt
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 784245799
+    num_examples: 505391
+  - name: test
+    num_bytes: 19997243
+    num_examples: 18433
+  download_size: 259222454
+  dataset_size: 804243042
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+  - split: test
+    path: data/test-*
+---

data/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:31eab5810af85c62e19ffc11f20cb89b29745622607d6e72ed37335e44f20d54
+size 2987024

data/train-00000-of-00002.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc51516cae5318de9b575f8759981c04d03429499940383d20a9372525942bdd
+size 214721252

data/train-00001-of-00002.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57bd47805233284587d88575318e5ca2c3d7d0caf5ef2c5c1e248995a53da844
+size 41514178