README
LongWriter-Zero RL Data
🤗 [Model] • 📃 [Paper] • 💾 [Dataset Card]
LongWriter-Zero RL Data is designed for ultra-long text generation via reinforcement learning. The dataset consists of conversational queries paired with length-range tags, which specify the desired output span (measured in words or Chinese characters). These annotations are used to train the LongWriter-Zero model, enabling it to consistently generate passages exceeding 10,000 words.
PS: We also included some general QA query to improve the model’s generalization ability. You can remove them based on the label range [0, 14000].
Dataset at a Glance
| Field | Type | Description |
|---|---|---|
idx | int | Unique example identifier |
query | string | User instruction / prompt (English or Chinese) |
label | object | JSON dict {"range": [low, high]} denoting the target word‑count interval |
Happy long-form writing!
Similar repos
README
LongWriter-Zero RL Data
🤗 [Model] • 📃 [Paper] • 💾 [Dataset Card]
LongWriter-Zero RL Data is designed for ultra-long text generation via reinforcement learning. The dataset consists of conversational queries paired with length-range tags, which specify the desired output span (measured in words or Chinese characters). These annotations are used to train the LongWriter-Zero model, enabling it to consistently generate passages exceeding 10,000 words.
PS: We also included some general QA query to improve the model’s generalization ability. You can remove them based on the label range [0, 14000].
Dataset at a Glance
| Field | Type | Description |
|---|---|---|
idx | int | Unique example identifier |
query | string | User instruction / prompt (English or Chinese) |
label | object | JSON dict {"range": [low, high]} denoting the target word‑count interval |
Happy long-form writing!