Upload README.md with huggingface_hub

README.md

@@ -1,41 +1,5 @@
 ---
 license: agpl-3.0
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
-  - split: eval
-    path: data/eval-*
-dataset_info:
-  features:
-  - name: source
-    dtype: string
-  - name: prompt
-    dtype: string
-  - name: system
-    dtype: string
-  - name: messages
-    list:
-    - name: role
-      dtype: string
-    - name: content
-      dtype: string
-    - name: reasoning_content
-      dtype: string
-    - name: tool_calls
-      dtype: string
-  - name: tools
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 130708904
-    num_examples: 9836
-  - name: eval
-    num_bytes: 32690514
-    num_examples: 2460
-  download_size: 130827186
-  dataset_size: 163399418
 ---
 
 This dataset merges two sources into a single SFT-ready set with train/eval
@@ -50,8 +14,9 @@ distributed under AGPL-3.0.
 ## Split
 Both source datasets were loaded in full (no weighting/subsampling),
 normalized to a common `messages` (OpenAI-style, tool calls preserved)
-schema, concatenated, shuffled together (seed 42), and split 80% train /
-20% eval on the combined pool.
+schema, concatenated, filtered to drop rows over ~8192 tokens (estimated
+at 4 chars/token, not a real tokenizer), shuffled together (seed 42), and
+split 80% train / 20% eval on the combined pool.
 
 ## Schema
 - `source`: origin dataset name