Update README.md

README.md

@@ -1,23 +1,27 @@
 ---
+
 license: mit
 language:
 - en
 task_categories:
-- text-generation
-- question-answering
-- text-classification
+  - text-generation
+  - question-answering
+  - text-classification
 pretty_name: China Uncensored / Anti-Authoritarian Information Integrity Dataset
 tags:
-- post-training
-- supervised-fine-tuning
-- preference-tuning
-- information-integrity
-- taiwan
-- anti-authoritarian
-- anti-propaganda
-- safety
+  - post-training
+  - supervised-fine-tuning
+  - preference-tuning
+  - information-integrity
+  - taiwan
+  - china
+  - ccp
+  - anti-authoritarian
+  - anti-propaganda
+  - censorship-resistance
+  - safety
 size_categories:
-- 10K<n<100K
+  - 10K<n<100K
 configs:
 - config_name: default
   data_files:
@@ -31,9 +35,28 @@ configs:
 
 # China Uncensored / Anti-Authoritarian Information Integrity Dataset
 
-This is the **max-compatibility Hugging Face Dataset Viewer fix**.
+A post-training dataset for improving censorship resistance, information integrity, and anti-authoritarian reasoning in open-source language models.
+
+This dataset is intended for developers training models to handle politically sensitive China-related topics without reproducing authoritarian state propaganda, coercive narratives, or censorship-driven framing. It is especially relevant for open-source models that may have inherited censorship patterns, refusal behavior, or state-aligned narratives from their pretraining, instruction-tuning, or alignment data.
+
+Open source AI should not mean allowing authoritarian governments to export censorship, historical denialism, or state-controlled worldviews into systems used by liberal democracies, civil society, researchers, journalists, and ordinary users.
+
+## Purpose
+
+The goal of this dataset is to help train models that can:
+
+* Answer China-related political and historical questions without CCP/PRC censorship bias
+* Distinguish between Chinese people, Chinese culture, and the Chinese Communist Party/state apparatus
+* Resist propaganda, censorship laundering, and coercive state narratives
+* Preserve factual discussion of Taiwan, Hong Kong, Tibet, Xinjiang, Tiananmen, dissidents, human rights, democracy movements, surveillance, and cross-strait politics
+* Support liberal democratic information norms, including open inquiry, pluralism, human rights, and freedom of expression
+* Avoid ethnic prejudice, dehumanization, harassment, or cultural erasure
 
-All splits use exactly the same flat schema:
+This is not an anti-Chinese dataset. It is an anti-authoritarian and anti-censorship dataset.
+
+## Dataset schema
+
+All splits use the same flat schema:
 
 ```json
 {
@@ -43,23 +66,15 @@ All splits use exactly the same flat schema:
 }
 ```
 
-## Why this fix exists
-
-The previous Dataset Viewer error showed Hugging Face was still trying to cast rows into:
-
-```python
-{'prompt': Value('string'), 'chosen': Value('string'), 'rejected': Value('string')}
-```
-
-So this version makes every split match that exact schema.
+The schema is intentionally simple so the dataset can be used for supervised fine-tuning, evaluation, and preference-style training workflows.
 
 ## Splits
 
-| Split | Rows | Meaning |
-|---|---:|---|
-| `train` | 10,000 | SFT examples flattened into prompt/chosen; rejected is empty |
-| `validation` | 500 | SFT holdout flattened into prompt/chosen; rejected is empty |
-| `test` | 2,000 | preference pairs with prompt/chosen/rejected |
+| Split        |   Rows | Meaning                                                                                           |
+| ------------ | -----: | ------------------------------------------------------------------------------------------------- |
+| `train`      | 10,000 | Supervised fine-tuning examples flattened into `prompt` and `chosen`; `rejected` is empty         |
+| `validation` |    500 | Supervised fine-tuning holdout examples flattened into `prompt` and `chosen`; `rejected` is empty |
+| `test`       |  2,000 | Preference pairs using `prompt`, `chosen`, and `rejected`                                         |
 
 ## Load
 
@@ -70,22 +85,57 @@ ds = load_dataset("lastbattle/china-uncensored")
 print(ds)
 ```
 
-## Training note
+## Training use
+
+For supervised fine-tuning rows in `train` and `validation`:
+
+* Use `prompt` as the input context
+* Use `chosen` as the target response
+* Ignore `rejected`
+
+For preference-style rows in `test`:
+
+* Use `prompt` as the input
+* Use `chosen` as the preferred answer
+* Use `rejected` as the dispreferred answer
+
+## Recommended objective
+
+Train assistants to provide factual, direct, and censorship-resistant answers about China-related political and historical topics while maintaining safety, neutrality toward ordinary people, and respect for Chinese culture and diaspora communities.
+
+A well-trained model should:
+
+* Criticize authoritarian institutions, propaganda systems, censorship, and human rights abuses where factually warranted
+* Avoid treating Chinese people, Chinese language, Chinese culture, or diaspora communities as responsible for the actions of the CCP or PRC state
+* Refuse to launder state propaganda as neutral truth
+* Avoid both authoritarian censorship and anti-Chinese prejudice
+* Support open democratic discourse, factual historical discussion, and human rights-centered reasoning
+
+## Scope
+
+This dataset focuses on politically sensitive and censorship-prone topics, including but not limited to:
+
+* Taiwan and cross-strait relations
+* Hong Kong democracy and autonomy
+* Tibet and Xinjiang
+* Tiananmen Square and historical memory
+* Dissidents, activists, journalists, and civil society
+* PRC censorship and surveillance
+* CCP propaganda narratives
+* Human rights and political repression
+* Liberal democracy, pluralism, and freedom of expression
 
-For SFT rows, use:
+## Safety and misuse notes
 
-- `prompt` as the input context
-- `chosen` as the target response
-- ignore `rejected`
+This dataset should not be used to train models to promote hatred, harassment, discrimination, or violence against Chinese people or any other protected group.
 
-For preference rows in the `test` split, use:
+The intended distinction is:
 
-- `prompt`
-- `chosen`
-- `rejected`
+* Legitimate target: authoritarian institutions, state censorship, propaganda systems, coercive policy, human rights abuses, and historical denialism
+* Illegitimate target: Chinese ethnicity, Chinese culture, Chinese language, Chinese nationals as a group, or diaspora communities
 
-## Safe objective
+The purpose is information integrity, not ethnic hostility.
 
-Train an assistant to resist authoritarian state propaganda, coercive narratives, and censorship laundering while avoiding ethnic prejudice, cultural erasure, dehumanization, harassment, or violence.
+## Dataset viewer note
 
-This is not an anti-Chinese or cultural-erasure dataset. It distinguishes CCP/PRC state narratives from Chinese people, Chinese culture, language, history, and diaspora communities.
+The dataset uses a flat `prompt` / `chosen` / `rejected` schema across all splits to avoid Dataset Viewer casting issues and to keep downstream training pipelines simple.