fda
Datasets
All datasets matching “fda”based-fdaThis dataset is adapted from the paper Language Models Enable Simple Systems for Generating
Structured Views of Heterogeneous Data Lakes. You can learn more about the data collection process there.
Please consider citing the following if you use this task in your work:
@article{arora2024simple,
title={Simple linear attention language models balance the recall-throughput tradeoff},
author={Arora, Simran and Eyuboglu, Sabri and Zhang, Michael and Timalsina, Aman and Alberti, Silas and… See the full description on the dataset page: https://huggingface.co/datasets/hazyresearch/based-fda.MP3
Still missing .json.gz Nov 20, 17:30
jkenne60 mwebb51
empty file
dwang58 pmoore34
Still missing .json.gz Nov 19, 14:00
mwebb51 npatton4 nvanflee sshres25
empty file
dwang58, mzg857
Broken
pmoore34
Still missing .json.gz Nov 15, 10:46
'hchen73','jkenne60','mwebb51','net','npatton4','nvanflee','rfranqui','sshres25'
Broken
'bfitzpa8','edayney', 'lhunte21','pmoore34','mzg857','sdasari7'
Still missing .json.gz Nov… See the full description on the dataset page: https://huggingface.co/datasets/fdac24/MP3.MP4
As of Nov 25, 20:00
Missing
jkenne60 mwebb51
No sections
ccanonac cwitt8 dwang58 ehead3 glapham harshvar hchen73 lhunte21 mzg857 nvanfleepmoore34 rfranqui tcatunca vgopu ygaikwad
In some cases you removed all newlines, so need to change regexp to extract section.
dwang58 mzg857 pmoore34
In the remaining cases newlines are there as \n, so there must be some other bug.
FDAbench-Full
v1.1 Update (2026-08-06) — multiple split
Strengthened the cross-source requirement that multiple-choice tasks are designed
around (selecting all correct options should require integrating both the SQL
result and the retrieved documents): 264 of 760 tasks were revised, with task IDs,
databases, and gold SQL unchanged. Documents-only accuracy drops from 61.5% to
38.7% while full-evidence accuracy stays at 80.6% (3 frontier models, strict
exact set match).
Diversified the number… See the full description on the dataset page: https://huggingface.co/datasets/FDAbench2026/FDAbench-Full.FDA_for_NLUFdabench-Lite
v1.1 Update (2026-08-06) — multiple split
Synced with FDABench-Full v1.1: all 49 multiple-choice tasks now carry the revised
question/option text and answer keys. The revision strengthens the cross-source
requirement (answering requires combining the SQL result with the retrieved
documents), diversifies the number of correct options, removes answer-count hints,
and balances option-style signals. Task IDs, databases, and gold SQL are unchanged.
See the FDABench-Full dataset card… See the full description on the dataset page: https://huggingface.co/datasets/FDAbench2026/Fdabench-Lite.