Expand/update data/val.jsonl

Expand/update data/train_split.jsonl

Expand/update data/train.jsonl

Update val split

Update train split

Expand: 200 -> 224 conversations

Fix: correct val split (stratified 90/10 from full 200, seed 42)

Fix: correct train split (stratified 90/10 from full 200, seed 42)

Move 90% split to train_split.jsonl (avoids clobbering train.jsonl)

Fix: restore full train.jsonl (200 records, renumbered IDs)

Add val split (90/10 stratified, seed 42)

Add train split (90/10 stratified, seed 42)

Add Indonesian Chat Conversations dataset