Update README.md

README.md

@@ -84,31 +84,6 @@ print(dataset)
 
 # Inspect column names in the “train” split
 print(dataset["train"].column_names)
-```
-
-By default, the single CSV is assigned a “train” split. You can then filter or sample by `context_range`:
-
-```python
-# Example: count examples per bucket
-from collections import Counter
-counts = Counter(dataset["train"]["context_range"])
-print(counts)  # e.g., {'3k': 100, '4k': 100, '8k': 100, '16k': 100, '32k': 100}
-
-# Filter to only 16k‐token contexts
-ds_16k = dataset["train"].filter(lambda x: x["context_range"] == "16k")
-print(len(ds_16k))
-```
-
-If you have downloaded `longbench_all_buckets_100.csv` locally, you can also load it via:
-
-```python
-from datasets import load_dataset
-
-dataset = load_dataset("csv", data_files="longbench_all_buckets_100.csv")
-print(dataset["train"].column_names)
-```
-
-After loading, each row will expose the fields:
 
 ```
 ["context", "question", "answer", "length", "dataset", "context_range"]