* How it confuses: Test-takers might select an option related to a different part of the conversation or lecture

Topic Shift Confusion * Example: The speaker may shift from one topic to another, options might include information from both topics

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

browse 1 citing papers

representative citing papers

eess.AS · 2024-10-24 · conditional · novelty 6.0

MMAU benchmark shows top audio-language models achieve only 52-53 percent accuracy on complex multi-task reasoning across 27 skills.

Showing 1 of 1 citing paper.

MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmark eess.AS · 2024-10-24 · conditional · none · ref 21
MMAU benchmark shows top audio-language models achieve only 52-53 percent accuracy on complex multi-task reasoning across 27 skills.