Update README.md

README.md

@@ -7,7 +7,7 @@ size_categories:
 task_categories:
 - conversational
 - text-generation
-pretty_name: UltraChat200k
+pretty_name: UltraChat 200k
 configs:
 - config_name: default
   data_files:
@@ -48,37 +48,31 @@ dataset_info:
   dataset_size: 3047427114
 ---
 
-# Dataset Card for UltraChat200k
+# Dataset Card for UltraChat 200k
 
 ## Dataset Description
 
-This is a pre-processed Supervised Fine-Tuning dataset used for training Zephyr-7b-beta, a state of the art 7b chat model.
+This is a heavily filtered version of the [UltraChat](https://github.com/thunlp/UltraChat) dataset and was used to train [Zephyr-7B-β](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta), a state of the art 7b chat model.
 
-The Zephyr-beta model is the best in class 7b model on three well known benchmarks:
+The original datasets consists of 1.4M dialogues generated by ChatGPT and spanning a wide range of topics. To create `UltraChat 200k`, we applied the following logic:
-- [MT Bench](https://huggingface.co/spaces/lmsys/mt-bench) - A multi-turn question set that uses GPT4 as a judge.
-- [Alpaca eval](https://tatsu-lab.github.io/alpaca_eval/) - An LLM-based automatic evaluation that is fast, cheap, and reliable. That tests the ability of models to follow general user instructions.
-- [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard) which aims to track, rank and evaluate open LLMs and chatbots.
 
-You can learn more about the techniques used to train Zephyr in the [Hugging Face Alignment Handbook](https://github.com/huggingface/alignment-handbook).
-
-
-The base dataset is [UltraChat](https://github.com/thunlp/UltraChat): an open-source, large-scale, and multi-round dialogue dataset.
-
-The dataset contains:
-- 🌏 **Questions about the World**: The dialogue data in this sector is derived from a wide range of inquiries related to concepts, entities, and objects from the real world. The topics covered are extensive, spanning areas such as technology, art, and entrepreneurship.
-- ✍🏻 **Writing and Creation**: The dialogue data in this sector is driven by the demands for writing/creation from scratch, and encompasses any tasks that an AI assistant may aid within the creative process, spanning from email composition to crafting narratives and plays, and beyond.
-- 📋 **Assistance on Existent Materials**: The dialogue data in this sector is generated based on existing materials, including but not limited to rewriting, continuation, summarization, and inference, covering a diverse range of topics.
-
-The following preprocessing was applied:
 - Selection of a subset of data for faster supervised fine tuning.
-- Truecasing of the dataset, as we observed around 5% of the data contained grammatical errors.
+- Truecasing of the dataset, as we observed around 5% of the data contained grammatical errors like "Hello. how are you?" instead of "Hello. How are you?"
-- Removal of dialogues where the assistant replies "I do not have emotions", "I don't have opinions"
+- Removal of dialogues where the assistant replies with phrases like "I do not have emotions" or "I don't have opinions", even for fact-based prompts that don't involve either.
 
 ## Dataset Structure
 
-The dataset contains two splits:
+The dataset has four splits, suitable for:
-- train - containing 207,865 examples
+
-- test - 23,110 examples
+* Supervised fine-tuning (`sft`).
+* Generation ranking (`gen`) via techniques like rejection sampling or PPO.
+
+The number of examples per split is shown as follows:
+
+
+|  train_sft | test_sft  | train_gen | test_gen |
+|:-------:|:-----------:|:-----:| :-----:|
+|  207865 |       23110 | 256032 | 28304 |
 
 The dataset is stored in parquet format with each entry using the following schema:
 ```