Benchmark

Qasper

QASPER is a dataset of 5,049 information-seeking questions and answers anchored in 1,585 NLP research papers. Questions are written by NLP practitioners…

Modality
text
Categories
long_context, reasoning
Openness
unknown
Source
llm_stats
Reported scores
2

Reported scores

llm_stats

ModelOrganizationReported valueReportedEvidence
Phi-3.5-mini-instructMicrosoft0.4192024-08-23evidence
Phi-3.5-MoE-instructMicrosoft0.42024-08-23evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.