woog/arena-prose-100-49-models

237 downloads in 30 days1 likesLast modified 23 Sep 2026View on Hugging Face

The checks could not identify a recognised licence ID in the dataset's metadata.

Declared licence: other, from the tagEvidence: README.md at 66298b5
Licence name on the card: mixed-source-and-model-output-terms, link: LICENSE.mdEvidence: README.md at 66298b5
Licence files: LICENSE.md, detected as cc-by-4.0Evidence: LICENSE.md at 66298b5
Declared sources: None listed in the card metadata
Access gated on Hugging Face: No

Each cell links to what was checked: a file, a file list or a Hub record.

Tasks

Citation text from the card

Prompts come from [LMSYS Chatbot Arena Conversations](https://huggingface.co/datasets/lmsys/chatbot_arena_conversations), revision `1b6335d42a1d2c7e34870c905d03ab964f7f2bd8`. Only opening user requests are used; archived assistant responses and votes are not included. The source contains 33,000 pairs collected in April–June 2023. Exact-agreement opening requests were deduplicated using NFC and whitespace normalization, producing 26,679 unique requests. Original wording was preserved for generation.

Candidates were ordered globally by SHA-256 of `pangram4-arena20-v1`, a NUL separator, and the normalized prompt ID. A single assistant reviewer screened the contiguous prefix of 320 candidates; the first 100 eligible prompts were retained. Eligible requests are self-contained English requests for substantial original prose. Coding, mathematical problem-solving, multiple-choice, rewriting/editing, translation, extraction, copied-text requests and short factual answers were excluded. No topic quotas were imposed. Category counts: 44 explanatory, 15 creative, 11 argumentative, 26 practical, 4 professional.

The original 20 selected prompts were retained when expanding to 100. The 80 additions were screened before their generation; outcomes from the original 20 had already been observed. All generators received the identical frozen sample. Reused successful generations were retained only when the full request matched. Responses were not replaced for refusal, quality, length, truncati

All datasets marked Unclear | JSON

The same facts for your agent

A free MCP server at mcp.datalicense.uk/mcp. Connect it from an MCP client such as Claude or Cursor.

dataset_licence({ "id": "woog/arena-prose-100-49-models" })
signal: "unclear"
stated_licence: "other"
sources: 0