# Hosted APIs

{% hint style="info" %}
These models are intended for use in frontend applications. For Python data analysis and backends, you should try our [PyPI package](/rumormill/toolkit/python-package.md), which runs locally.
{% endhint %}

### [fairlyAspects](https://huggingface.co/maximuspowers/bias-type-classifier) (types-of bias)

Endpoint: [`https://t41xs75wejr14zht.us-east-1.aws.endpoints.huggingface.cloud`](https://t41xs75wejr14zht.us-east-1.aws.endpoints.huggingface.cloud)

{% embed url="<https://huggingface.co/maximuspowers/bias-type-classifier>" %}

{% tabs %}
{% tab title="JavaScript" %}

```javascript
async function query(data) {
	const response = await fetch(
		"https://t41xs75wejr14zht.us-east-1.aws.endpoints.huggingface.cloud",
		{
			headers: { 
				"Accept" : "application/json",
				"Content-Type": "application/json" 
			},
			method: "POST",
			body: JSON.stringify(data),
		}
	);
	const result = await response.json();
	return result;
}

query({
    "inputs": "Your sentence for bias aspect classification",
    "parameters": {
        "top_k": 3,
        "function_to_apply": "sigmoid"
    }
}).then((response) => {
	console.log(JSON.stringify(response));
});
```

#### All parameters are optional

* `top_k`: Number of classes and scores to return. Defaults to return all 11 classes and their scores.
* `function_to_apply`: Activation function to use on the model outputs. Can be `"sigmoid"`(multi-label), `"softmax"` (mult-class), or `None` for raw logits. Defaults to `"softmax"`.
  {% endtab %}

{% tab title="Python" %}

```python
import requests

API_URL = "https://t41xs75wejr14zht.us-east-1.aws.endpoints.huggingface.cloud"
headers = {
	"Accept" : "application/json",
	"Content-Type": "application/json" 
}

def query(payload):
	response = requests.post(API_URL, headers=headers, json=payload)
	return response.json()

output = query({
	"inputs": "Your sentence for bias aspect classification",
	"parameters": {
		"top_k": 3,
		"function_to_apply": "sigmoid"
	}
})
```

#### All parameters are optional

* `top_k`: Number of classes and scores to return. Defaults to return all 11 classes and their scores.
* `function_to_apply`: Activation function to use on the model outputs. Can be `"sigmoid"`(multi-label), `"softmax"` (mult-class), or `None` for raw logits. Defaults to `"softmax"`.
  {% endtab %}

{% tab title="cURL" %}

<pre class="language-bash"><code class="lang-bash"><strong>curl "https://t41xs75wejr14zht.us-east-1.aws.endpoints.huggingface.cloud" \
</strong>-X POST \
-H "Accept: application/json" \-H "Content-Type: application/json" \
-d '{
    "inputs": "Your sentence for bias aspect classification",
    "parameters": {
        "top_k": 3,
        "function_to_apply": "sigmoid"
    }
}'
</code></pre>

#### All parameters are optional

* `top_k`: Number of classes and scores to return. Defaults to return all 11 classes and their scores.
* `function_to_apply`: Activation function to use on the model outputs. Can be `"sigmoid"`(multi-label), `"softmax"` (mult-class), or `None` for raw logits. Defaults to `"softmax"`.
  {% endtab %}
  {% endtabs %}

***

### [Binary Bias Classification](/rumormill/research/sequence-classification/binary.md)

Endpoint: [`https://e29pozks7ptdl5gw.us-east-1.aws.endpoints.huggingface.cloud`](https://e29pozks7ptdl5gw.us-east-1.aws.endpoints.huggingface.cloud)

{% embed url="<https://huggingface.co/valurank/distilroberta-bias>" %}
Similar architecture to [Dbias](https://arxiv.org/abs/2208.05777).
{% endembed %}

{% tabs %}
{% tab title="JavaScript" %}

```javascript
async function query(data) {
	const response = await fetch(
		"https://e29pozks7ptdl5gw.us-east-1.aws.endpoints.huggingface.cloud",
		{
			headers: { 
				"Accept" : "application/json",
				"Content-Type": "application/json" 
			},
			method: "POST",
			body: JSON.stringify(data),
		}
	);
	const result = await response.json();
	return result;
}

query({
    "inputs": "Your sentence for binary bias classification",
    "parameters": {
        "top_k": 2,
        "function_to_apply": "sigmoid"
    }
}).then((response) => {
	console.log(JSON.stringify(response));
});
```

#### All parameters are optional

* `top_k`: Can be `1` or `2`, defaults to `1`, meaning it just returns the highest probable class (Biased or Neutral).
* `function_to_apply`: Activation function to use on the model outputs. Can be `"sigmoid"`(multi-label), `"softmax"` (mult-class), or `None` for raw logits. Defaults to `"softmax"`.
  {% endtab %}

{% tab title="Python" %}

```python
import requests
```

#### All parameters are optional

* `top_k`: Can be `1` or `2`, defaults to `1`, meaning it just returns the highest probable class (Biased or Neutral).
* `function_to_apply`: Activation function to use on the model outputs. Can be `"sigmoid"`(multi-label), `"softmax"` (mult-class), or `None` for raw logits. Defaults to `"softmax"`.
  {% endtab %}

{% tab title="cURL" %}

```bash
curl "https://e29pozks7ptdl5gw.us-east-1.aws.endpoints.huggingface.cloud" \
-X POST \
-H "Accept: application/json" \-H "Content-Type: application/json" \
-d '{
    "inputs": "Your sentence for binary bias classification",
    "parameters": {
        "top_k": 2,
        "function_to_apply": "sigmoid"
    }
}'
```

#### All parameters are optional

* `top_k`: Can be `1` or `2`, defaults to `1`, meaning it just returns the highest probable class (Biased or Neutral).
* `function_to_apply`: Activation function to use on the model outputs. Can be `"sigmoid"`(multi-label), `"softmax"` (mult-class), or `None` for raw logits. Defaults to `"softmax"`.
  {% endtab %}
  {% endtabs %}

***

### [Token Classification](/rumormill/research/ner/token-classification.md) (of generalizations, unfairness, and stereotypes)

Endpoint: [`https://e29pozks7ptdl5gw.us-east-1.aws.endpoints.huggingface.cloud`](https://e29pozks7ptdl5gw.us-east-1.aws.endpoints.huggingface.cloud)

{% embed url="<https://huggingface.co/ethical-spectacle/social-bias-ner>" %}
[GUS-Net](https://arxiv.org/abs/2410.08388) Model
{% endembed %}

{% tabs %}
{% tab title="JavaScript" %}

```javascript
async function query(data) {
	const response = await fetch(
		"https://mo5fr3ll9qufbwuy.us-east-1.aws.endpoints.huggingface.cloud",
		{
			headers: { 
				"Accept" : "application/json",
				"Content-Type": "application/json" 
			},
			method: "POST",
			body: JSON.stringify(data),
		}
	);
	const result = await response.json();
	return result;
}

query({
    "inputs": "Hello world!",
    "parameters": {}
}).then((response) => {
	console.log(JSON.stringify(response));
});
```

\**No params needed*\*
{% endtab %}

{% tab title="Python" %}

```python
import requests

API_URL = "https://e29pozks7ptdl5gw.us-east-1.aws.endpoints.huggingface.cloud"
headers = {
	"Accept" : "application/json",
	"Content-Type": "application/json" 
}

def query(payload):
	response = requests.post(API_URL, headers=headers, json=payload)
	return response.json()

output = query({
	"inputs": "I like you. I love you",
	"parameters": {
		"top_k": 2,
		"function_to_apply": "sigmoid"
	}
})
```

\**No params needed*\*
{% endtab %}

{% tab title="cURL" %}

```bash
curl "https://e29pozks7ptdl5gw.us-east-1.aws.endpoints.huggingface.cloud" \
-X POST \
-H "Accept: application/json" \-H "Content-Type: application/json" \
-d '{
    "inputs": "I like you. I love you",
    "parameters": {
        "top_k": 2,
        "function_to_apply": "sigmoid"
    }
}'
```

\**No params needed*\*
{% endtab %}
{% endtabs %}

### Example Response:

```json
[
  {
    "token": "data",
    "labels": [
      "B-STEREO",
      "B-GEN"
    ]
  },
  {
    "token": "scientists",
    "labels": [
      "B-GEN",
      "I-GEN"
    ]
  },
  {
    "token": "are",
    "labels": [
      "I-STEREO"
    ]
  },
  {
    "token": "so",
    "labels": [
      "I-STEREO"
    ]
  },
  {
    "token": "smart",
    "labels": [
      "I-STEREO"
    ]
  }
]
```


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://ethical-spectacle-research.gitbook.io/rumormill/toolkit/hosted-apis.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.