nikolina-p/intention-request-to-talk-to

---
dataset_info:
features:
- name: text
dtype: string
- name: labels
dtype:
class_label:
names:
'0': '0'
'1': '1'
splits:
- name: train
num_bytes: 22324
num_examples: 542
- name: validation
num_bytes: 2512
num_examples: 61
- name: test
num_bytes: 6156
num_examples: 141
download_size: 16470
dataset_size: 30992
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
license: mit
task_categories:
- text-classification
language:
- en
tags:
- intent-classification
pretty_name: Intention Request to Talk To
size_categories:
- n<1K
---

# Intention Request to Talk To

## Dataset Description

This dataset is a binary text-classification dataset for detecting whether a user is requesting to speak with a specific named person.

### Supported Tasks
- Binary text classification
- Intent detection
- Chatbot routing / access-control prototypes

### Languages
- English

### Data Fields

- `text`: user input text
- `labels`:
- `0` = not a request to speak to a specific named person
- `1` = polite request to speak to a specific named person

### Data Instances

Example positive instance:

```json
{"text": "May I speak with Richard?", "labels": 1}
{"text": "Hello, Sarah mentioned I should talk to Michael for details. Could you put me through?", "labels": 1}
```

Example negative instance:

```json
{"text": "Who is Richard?", "labels": 0}
{"text": "Can I talk to her?", "labels": 0}
{"text": "Is anyone available to talk?", "labels": 0}
{"text": "They told me to ask to talk to Susan. Who is she?", "labels": 0}
```

### Dataset Creation

This dataset was created for a small intent-classification project involving chatbot authentication/routing behavior.

Positive examples are short user utterances that politely request to speak with a named person.

Negative examples include:

- general questions about a person
- unrelated chat messages
- statements
- questions not asking to speak to someone
- utterances without a specific named target

### Limitations
- Short-text focused
- English only
- Small scope
- Binary framing may not capture ambiguous requests