[Bug Report] PretokenizeRunner fails with streaming=True

[keltin13]

Describe the bug
PretokenizeRunner throws an error when streaming=True because num_proc is not a valid argument for an IterableDataset.

Code example
Adding the following test to tests/training/test_pretokenize_runner will raise an error:

def test_pretokenize_runner_streaming_dataset():
    cfg = PretokenizeRunnerConfig(
        tokenizer_name="gpt2",
        context_size=10,
        num_proc=2,
        dataset_path="NeelNanda/c4-10k",
        split="train",
        streaming=True,
    )
    dataset = PretokenizeRunner(cfg).run()

>       tokenized_dataset = dataset.map(
            process_examples,
            batched=True,
            batch_size=cfg.pretokenize_batch_size,
            num_proc=cfg.num_proc,
            remove_columns=dataset.column_names,
        )
E       TypeError: IterableDataset.map() got an unexpected keyword argument 'num_proc'

sae_lens/pretokenize_runner.py:102: TypeError

System Info
Describe the characteristic of your environment:

• Installed SAELens from source, as described in CONTRIBUTING.md
• MacOS
• Python 3.10

Checklist

• I have checked that there is no similar issue in the repo (required)