Datasets:
Dataset card: declare column types for the two price configurations
Browse files
README.md
CHANGED
|
@@ -41,8 +41,44 @@ configs:
|
|
| 41 |
data_files: manifold/markets_summary.csv
|
| 42 |
- config_name: polymarket_prices
|
| 43 |
data_files: polymarket/prices/*.csv
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 44 |
- config_name: manifold_prices
|
| 45 |
data_files: manifold/prices/*.csv
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 46 |
---
|
| 47 |
|
| 48 |
# StakeBench: Evaluating Language Understanding Grounded in Market Commitment
|
|
@@ -77,7 +113,7 @@ The key design principle is to replace perception-based supervision with **revea
|
|
| 77 |
|
| 78 |
## Hugging Face notes
|
| 79 |
|
| 80 |
-
- **Loading.** Each main CSV is exposed as a configuration of this dataset (see the YAML header): `polymarket_comments` (default), `manifold_comments`, `polymarket_positions`, `manifold_positions`, `polymarket_markets`, `manifold_markets`, `polymarket_topics`, `manifold_markets_summary`, `polymarket_prices`, `manifold_prices`. The `raw/` and `model_outputs/` folders are plain files; download them with `snapshot_download`.
|
| 81 |
|
| 82 |
```python
|
| 83 |
from datasets import load_dataset
|
|
|
|
| 41 |
data_files: manifold/markets_summary.csv
|
| 42 |
- config_name: polymarket_prices
|
| 43 |
data_files: polymarket/prices/*.csv
|
| 44 |
+
features:
|
| 45 |
+
- name: market_id
|
| 46 |
+
dtype: int64
|
| 47 |
+
- name: slug
|
| 48 |
+
dtype: string
|
| 49 |
+
- name: question
|
| 50 |
+
dtype: string
|
| 51 |
+
- name: category
|
| 52 |
+
dtype: string
|
| 53 |
+
- name: token_id
|
| 54 |
+
dtype: string
|
| 55 |
+
- name: outcome
|
| 56 |
+
dtype: string
|
| 57 |
+
- name: ts
|
| 58 |
+
dtype: int64
|
| 59 |
+
- name: datetime
|
| 60 |
+
dtype: string
|
| 61 |
+
- name: price
|
| 62 |
+
dtype: float64
|
| 63 |
- config_name: manifold_prices
|
| 64 |
data_files: manifold/prices/*.csv
|
| 65 |
+
features:
|
| 66 |
+
- name: market_id
|
| 67 |
+
dtype: string
|
| 68 |
+
- name: slug
|
| 69 |
+
dtype: string
|
| 70 |
+
- name: question
|
| 71 |
+
dtype: string
|
| 72 |
+
- name: category
|
| 73 |
+
dtype: string
|
| 74 |
+
- name: ts
|
| 75 |
+
dtype: int64
|
| 76 |
+
- name: datetime
|
| 77 |
+
dtype: string
|
| 78 |
+
- name: outcome
|
| 79 |
+
dtype: string
|
| 80 |
+
- name: price
|
| 81 |
+
dtype: float64
|
| 82 |
---
|
| 83 |
|
| 84 |
# StakeBench: Evaluating Language Understanding Grounded in Market Commitment
|
|
|
|
| 113 |
|
| 114 |
## Hugging Face notes
|
| 115 |
|
| 116 |
+
- **Loading.** Each main CSV is exposed as a configuration of this dataset (see the YAML header): `polymarket_comments` (default), `manifold_comments`, `polymarket_positions`, `manifold_positions`, `polymarket_markets`, `manifold_markets`, `polymarket_topics`, `manifold_markets_summary`, `polymarket_prices`, `manifold_prices`. The `raw/` and `model_outputs/` folders are plain files; download them with `snapshot_download`. The two price configurations declare their column types in the YAML header (`token_id` as a string, because Polymarket token IDs exceed 64-bit integers; `category` as a string, because it is empty in some files); when reading the CSV files directly, pass `dtype={"token_id": str}` to pandas.
|
| 117 |
|
| 118 |
```python
|
| 119 |
from datasets import load_dataset
|