Wikibench

Natural Questions

reasoning official site →

Multilingual: no

Natural Questions is a question answering dataset featuring real anonymized queries issued to Google search engine. It contains 307,373 training examples where annotators provide long answers (passages) and short answers (entities) from Wikipedia pages, or mark them as unanswerable.

Methodology

Imported from llm-stats public benchmark metadata. Modality: text. Max score: 1. Categories: general, reasoning, search. Language: en. Verified by llm-stats: no.

Leaderboard

Showing 7 of 7 results

  1. Gemma 2 27B self-reported llm-stats
    34.5%
  2. Mistral NeMo Instruct self-reported llm-stats
    31.2%
  3. Gemma 2 9B self-reported llm-stats
    29.2%
  4. Gemma 3n E4B self-reported llm-stats
    20.9%
  5. Gemma 3n E4B Instructed LiteRT Preview self-reported llm-stats
    20.9%
  6. Gemma 3n E2B self-reported llm-stats
    15.5%
  7. Gemma 3n E2B Instructed LiteRT (Preview) self-reported llm-stats
    15.5%