WildTrace Benchmark
Explore score history ↗WildTrace evaluates long-context reasoning over naturally occurring long-form sources, focusing on integrating evidence dispersed across distant passages. It includes 481 tasks over 214 sources with seven source-internal evidence geometries, with multi-stage validation ensuring answer groundedness and clue necessity.
- Released
- 2026-07-10
- Institutions
- Hong Kong University of Science and Technology · Alibaba Group (China)
- Evaluates
- General AI, Language & Knowledge, Long Context, Reasoning
- Openness
- unknown
- Importer
- Claire Radar
- Review status
- unreviewed
WildTrace paper
No reported scores are on record for this benchmark yet.