IFM/FannOrFlop-MCQ_verified

Fann or Flop: Arabic poetry multiple-choice tasks (verified)

A filtered version of multiple-choice questions on Arabic poetry derived from the poems of the Fann or Flop benchmark, with samples flagged by a multi-judge audit removed. There is one config per task (14 configs): poem_text__genre, poem_text__meter, poem_text__poet_name, poet_name__genre, and so on, where the names before and after __ describe what is given in the question and what is asked.

What "verified" means. Samples were removed when all three LLM judges flagged a significant defect. This is not a human review, and only samples that baseline models answered incorrectly were audited, so some defects may remain in the kept data.

Source

• The poems and their metadata (poet, era, genre, meter) come from omkarthawakar/FannOrFlop (Apache-2.0), revision 874b3f82d066e786369158a8ca3967585712486e (6,984 poems with explanations).
• The upstream dataset contains no questions, options or answer keys. The multiple-choice questions here (6,984 rows across the 14 tasks, the 50_verses version, which uses up to 50 verses of a poem) were generated from those poems for evaluation, and the generation script with its distractors is not available. For 5,719 of the 6,984 questions the verses shown were located in the upstream poems, and the answers checked against the upstream genre, meter, era and poet metadata (about 2,960 answers checked, no disagreement). Questions that give only a poet's name have no verses to locate.
• This is therefore a derived dataset, not the original Fann or Flop release.

Columns

question (prompt text including the options), options (option letters) and correct_answer (gold letter). Tasks that include a poem show its verses in the question.

Audit and filtering

Samples were screened with the multi-judge audit described in IFM/discarded_samples (revision 3a3c1339, config FannOrFlop_MCQ). Three language-model judges (GPT-5.6 Sol, Gemini 3.8 Flash, Claude Opus 5) assessed candidate samples that strong baselines answered incorrectly. A sample is excluded only when all three judges report a significant issue (a problem with the question, the options or the gold label that prevents reliable answering or scoring). Excluded samples were matched to this benchmark by the row index (doc_id) assigned by lm-evaluation-harness, which equals the 0-based row index within that task's question file in the evaluation copy (the 14 configs here; the upstream poem file has no questions); every matched row was additionally checked against the discarded record's question text and gold label. All other rows are kept in their original order, with the column names and values of the evaluation copy.

Config	Original rows	Excluded	Kept
genre__poem_text__poet_name__meter	200	1	199
poem_text__genre	550	36	514
poem_text__keywords	621	2	619
poem_text__meter	745	1	744
poem_text__poem_title	876	0	876
poem_text__poet_era	638	4	634
poem_text__poet_name	1042	0	1042
poem_text__poet_name__genre	321	28	293
poem_text__poet_name__meter	284	1	283
poem_text__poet_name__poet_era	225	12	213
poem_text__poet_name__rhyme	226	7	219
poet_name__genre	416	102	314
poet_name__meter	364	12	352
poet_name__poet_era	476	24	452
Total	6984	230	6754

License

The poems come from an Apache-2.0 dataset, which allows derived works with attribution. Please keep the attribution below.

Citation

Please cite the original Fann or Flop authors:

@misc{alghallabi2025fannflopmultigenremultiera,
  title         = {Fann or Flop: A Multigenre, Multiera Benchmark for Arabic Poetry Understanding in LLMs},
  author        = {Wafa Alghallabi and Ritesh Thawkar and Sara Ghaboura and Ketan More and Omkar Thawakar and Hisham Cholakkal and Salman Khan and Rao Muhammad Anwer},
  year          = {2025},
  eprint        = {2505.18152},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  url           = {https://arxiv.org/abs/2505.18152}
}

@misc{jais3_2026,
  title  = {Jais 3: Introducing Arabic-centric Agentic AI},
  author = {{IFM Team, Inception42 Team, and Cerebras Team}},
  year   = {2026},
  url    = {https://ifm.ai/blog/jais3/},
}