MMAU benchmark shows top audio-language models achieve only 52-53 percent accuracy on complex multi-task reasoning across 27 skills.
* How it confuses: Test-takers might select an option related to a different part of the conversation or lecture
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.AS 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmark
MMAU benchmark shows top audio-language models achieve only 52-53 percent accuracy on complex multi-task reasoning across 27 skills.