[Bug] token-classification return start=null and end=null instead of index in original text

**Describe the bug**
*A clear and concise description of what the bug is.*

**How to reproduce**
*Steps or a minimal working example to reproduce the behavior*
```js
let classifier = await pipeline('token-classification', 'Xenova/bert-base-NER');
let output = await classifier('My name is Sarah and I live in London');

[
{
"entity": "B-PER",
"score": 0.9980159401893616,
"index": 4,
"word": "Sarah",
"start": null,
"end": null
},
{
"entity": "B-LOC",
"score": 0.9994593262672424,
"index": 9,
"word": "London",
"start": null,
"end": null
}
]
```
**Expected behavior**

I expected "start" and "end" to point to the start and end in the original text.

**Environment**
- Transformers.js version: https://cdn.jsdelivr.net/npm/@xenova/transformers@2.6.2
- Browser (if applicable): Chrome
- Operating system (if applicable): MacOS
- Other:

**Additional context**

The documentation for the python version gives back proper start and end points.
https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.TokenClassificationPipeline.example