ElephantBench
Explore score history ↗Closed-book knowledge probe with 1,094 source-traceable questions that measure whether language models recall multiple verified accounts of long-tail facts, scored via complete, partial, or failed recall rates.
- Released
- 2026-08-28
- Institutions
- Tencent (China) · Tsinghua University · University of Warwick
- Evaluates
- Factuality, General AI, Language & Knowledge, cs.CL
- Openness
- unknown
- Importer
- Claire Radar
- Review status
- ai-reviewed
ElephantBench paper and code
No reported scores are on record for this benchmark yet.