luoojason/mm-long-storytelling-bench

MM Long Storytelling Bench Long-context reading comprehension over whole public-domain novels. Each question needs the full book (or, for cross-novel items, two to five books, either unrelated novels or books from one series) to answer. 1,749 questions · 146 novels · 329 distinct contexts · median input 103k tokens (max 812k). Token counts are o200k_base, the tokenizer tiktoken maps to gpt-5. Configs config rows what it is default 1,749 one row per… See the full description on the dataset page: https://huggingface.co/datasets/luoojason/mm-long-storytelling-bench.

<iframe src="https://silytics.si/embed/item/16113cb2-011d-4153-b97f-09373984370e" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.