ryankamiri/prolong-qwen

ProLong books tokenized for Qwen Source: princeton-nlp/prolong-data-64K, subset book-65536. Decoded with princeton-nlp/Llama-3-8B-ProLong-64k-Base and encoded with Qwen/Qwen3-1.7B-Base. Each source record is independent and truncated to at most 65,536 target tokens. Shorter records are unpadded. Source special tokens are removed; target BOS/EOS are not added. See manifest.json for pinned revisions, settings, and counts. Reset model session state between records. No tokenizer… See the full description on the dataset page: https://huggingface.co/datasets/ryankamiri/prolong-qwen.

<iframe src="https://silytics.si/embed/item/d85aedb1-658c-449c-b3f3-5937842a2ad7" 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 Information may be incomplete or out of date.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 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.