hear_voice_bias_benchmark
HEAR (Human-recorded Evaluation of Audio-LLM bias by Real speakers), a large-scale, ecologically valid benchmark comprising real human audio samples from demographically diverse participants.
2 stars
Python
Your first custom repo explanation is free. Reading existing public explainers always stays free.
This will take 10-20 minutes. You can close the tab and come back later.
More from facebookresearch
View all →
More from facebookresearch
View all →
Starting generation...
Checking for an available workflow slot.
Researching repository
Researching web & landscape
Searching X/Twitter
Planning the story
Writing the article
Generating illustrations
Building interactive diagrams
Adding expert commentary
Assembling final article
Publishing