Benchmark RadarAI BENCHMARK PROFILE
DAS-Bench
DAS-Bench provides 30 topics for evaluating academic survey generation, with DAS-Eval scoring 16 criteria across citation quality, taxonomic synthesis, hierarchical discourse, and manuscript reliability.
- Released
- 2026-08-18
- Readiness
- Runnable
- Primary field
- General AI
Why it matters
DAS-Bench introduces a standardized, multi-criteria evaluation for publication-oriented survey generation, enabling comparison across systems beyond existing ad hoc assessments.
Motivation
Academic surveys play a central role in organizing rapidly expanding scholarly literature, yet their construction requires extensive paper analysis, coherent knowledge organization, fine-grained citation support, and reliable manuscript assembly.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.