Dataset Viewer
Auto-converted to Parquet Duplicate
context
stringclasses
1 value
question
stringlengths
236
290
answer_prefix
stringclasses
1 value
answers
listlengths
1
9
task
stringclasses
1 value
max_new_tokens
int64
256
256
"You will be given a list of documents. You need to read carefully and understand all of them. Then (...TRUNCATED)
"====== Now let's start! ======\nBased on the documents above, can you answer the following query? P(...TRUNCATED)
Final Answer:
[ "the following day" ]
nq_128k
256
"You will be given a list of documents. You need to read carefully and understand all of them. Then (...TRUNCATED)
"====== Now let's start! ======\nBased on the documents above, can you answer the following query? P(...TRUNCATED)
Final Answer:
[ "Spain", "Taíno", "indigenous Taíno people" ]
nq_128k
256
"You will be given a list of documents. You need to read carefully and understand all of them. Then (...TRUNCATED)
"====== Now let's start! ======\nBased on the documents above, can you answer the following query? P(...TRUNCATED)
Final Answer:
[ "September 13, 2012" ]
nq_128k
256
"You will be given a list of documents. You need to read carefully and understand all of them. Then (...TRUNCATED)
"====== Now let's start! ======\nBased on the documents above, can you answer the following query? P(...TRUNCATED)
Final Answer:
[ "Haliaeetus" ]
nq_128k
256
"You will be given a list of documents. You need to read carefully and understand all of them. Then (...TRUNCATED)
"====== Now let's start! ======\nBased on the documents above, can you answer the following query? P(...TRUNCATED)
Final Answer:
[ "Kobol's Last Gleaming" ]
nq_128k
256
"You will be given a list of documents. You need to read carefully and understand all of them. Then (...TRUNCATED)
"====== Now let's start! ======\nBased on the documents above, can you answer the following query? P(...TRUNCATED)
Final Answer:
[ "Nala" ]
nq_128k
256
"You will be given a list of documents. You need to read carefully and understand all of them. Then (...TRUNCATED)
"====== Now let's start! ======\nBased on the documents above, can you answer the following query? P(...TRUNCATED)
Final Answer:
[ "Christopher Lloyd" ]
nq_128k
256
"You will be given a list of documents. You need to read carefully and understand all of them. Then (...TRUNCATED)
"====== Now let's start! ======\nBased on the documents above, can you answer the following query? P(...TRUNCATED)
Final Answer:
[ "Charlotte of Mecklenburg-Strelitz" ]
nq_128k
256
"You will be given a list of documents. You need to read carefully and understand all of them. Then (...TRUNCATED)
"====== Now let's start! ======\nBased on the documents above, can you answer the following query? P(...TRUNCATED)
Final Answer:
[ "Glenn Close" ]
nq_128k
256
"You will be given a list of documents. You need to read carefully and understand all of them. Then (...TRUNCATED)
"====== Now let's start! ======\nBased on the documents above, can you answer the following query? P(...TRUNCATED)
Final Answer:
[ "the Ramones" ]
nq_128k
256
End of preview. Expand in Data Studio

LOFT RAG - Natural Questions (128k)

Dataset Description

This dataset is part of the LOFT (Long-context Open Foundation Tasks) benchmark, specifically the RAG (Retrieval-Augmented Generation) task.

  • Dataset: Natural Questions
  • Context Length: 128k
  • Task Type: RAG (Retrieval-Augmented Generation)
  • Language: English
  • Source: LOFT Benchmark (Google DeepMind)

Dataset Structure

Data Fields

  • context (string): Full prompt context including corpus documents and few-shot examples
  • question (string): Query separator + query format + query text
  • answer_prefix (string): Prefix for answer generation ("Final Answer: ")
  • answers (list[string]): Ground truth answers
  • task (string): Task identifier (e.g., "nq_128k")
  • max_new_tokens (int64): Maximum tokens for generation (256)

Data Splits

  • dev: Development set (10 examples)
  • test: Test set (100 examples)

Usage

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("loft-rag-nq-128k")

# Access splits
dev_data = dataset["dev"]
df_dev = dev_data.to_pandas()
test_data = dataset["test"]
df_test = test_data.to_pandas()

# Example usage
sample = dataset["dev"][0] if "dev" in dataset else dataset["test"][0]
context = sample["context"]
question = sample["question"]
answers = sample["answers"]

Dataset Creation

This dataset was converted from LOFT's original format to HuggingFace format using exact LOFT prompt construction to ensure 100% fidelity.

  • Prompt Construction: Uses LOFT's PromptRegistry and concatenate_chunks() for exact prompt matching
  • Few-shot Examples: Preserved exactly as in LOFT (5 examples)
  • Corpus Documents: Full corpus included in context (corpus-in-context approach)
  • Verification: All prompts verified to match LOFT originals exactly

Related Datasets

All LOFT RAG datasets are available under the loft-rag-* namespace:

Citation

@article{{loft2024,
  title={{LOFT: Long-context Open Foundation Tasks}},
  author={{Google DeepMind}},
  year={{2024}},
  url={{https://github.com/google-deepmind/loft}}
}}

License

Apache 2.0

Downloads last month
37