BSVs dataset @ 16k seqlen, contiguous episodes with only model deemed high quality samples and the rewritten goals

README.md

@@ -0,0 +1,32 @@
+---
+dataset_info:
+  features:
+  - name: uid
+    dtype: string
+  - name: parent
+    dtype: string
+  - name: episode_id
+    dtype: int64
+  - name: step_id
+    dtype: int64
+  - name: messages
+    dtype: string
+  - name: tools
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 1720140890.4169824
+    num_examples: 71627
+  - name: test
+    num_bytes: 3962517.5830176063
+    num_examples: 165
+  download_size: 285558363
+  dataset_size: 1724103408.0
+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 sha256:40007f7b335c29ba65a69ece891af93c1d6705263f34f050d60342a96bb8be4c
+size 720296

data/train-00000-of-00004.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49b9a91a505ec7af4b9ede818c5f214570914fc6d45549fa2dee373c2e3e69c4
+size 71868415

data/train-00001-of-00004.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d892e0e8d0b68a330c79bd8da8bc80bc2d759827568016b354bf492f6b05be5b
+size 70719359

data/train-00002-of-00004.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd50bb31607230d5d79b0af8f6b49ca61e7833ee2d7b58f387e79cd11bdd3b90
+size 70919583

data/train-00003-of-00004.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd421633b5197684967076bd32e389fe3cd39a75e6d7100c8c24ccd4b03e405d
+size 71330710