README.md

---
pretty_name: Intent Detection Benchmark

task_categories:
- text-classification

language:
- en

tags:
- benchmark
- intent-detection
- search
- evaluation

size_categories:
- n<1K

model-index:
- name: GPT-4.1
  results:
  - task:
      type: text-classification
      name: Intent Detection
    dataset:
      name: Intent Detection Benchmark
      type: your-org/intent-detection-benchmark
    metrics:
    - type: accuracy
      value: 0.92
      name: Accuracy
    - type: f1
      value: 0.91
      name: F1

- name: Llama-3-70B
  results:
  - task:
      type: text-classification
      name: Intent Detection
    dataset:
      name: Intent Detection Benchmark
      type: your-org/intent-detection-benchmark
    metrics:
    - type: accuracy
      value: 0.87
      name: Accuracy
    - type: f1
      value: 0.86
      name: F1
---

# Intent Detection Benchmark

A benchmark dataset for e-commerce and search intent classification.

## Task

Given a query, predict the user intent.

Example intents:
- product_search
- returns
- order_tracking
- subscription_cancel
- purchase_accessories

---

# Leaderboard

| Model | Accuracy | F1 | Latency |
|---|---|---|---|
| GPT-4.1 | 0.92 | 0.91 | 820ms |
| Llama-3-70B | 0.87 | 0.86 | 410ms |

---

# Dataset Format

| query | intent |
|---|---|
| buy iphone charger | purchase_accessories |

---

# Citation

```bibtex
@dataset{intent_detection_benchmark,
  title={Intent Detection Benchmark},
  year={2026}
}
```