Benchmark RadarAI BENCHMARK PROFILE
EEG-EditBench
EEG-EditBench evaluates EEG-to-image retrieval models using 2,137 controlled edits of 200 THINGS-EEG2 test images, covering object identity, attributes, background, and presence, with metrics like 200-way accuracy and 2AFC accuracy.
- Released
- 2026-07-30
- Readiness
- Runnable
- Primary field
- General AI
Why it matters
Standard retrieval accuracy can mask whether models truly preserve visual information. EEG-EditBench provides a controlled, repeatable protocol for probing fine-grained visual distinctions, informing development of more robust EEG decoding models.
Motivation
Recent EEG-to-image retrieval models have achieved strong performance in identifying viewed images from semantically diverse candidates.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.