VCB-Bench

VCB-Bench: An Evaluation Benchmark for Audio-Grounded Large Language Model Conversational Agents Introduction Voice Chat Bot Bench (VCB Bench) is a high-quality Chinese benchmark built entirely on real human speech. It evaluates large audio language models (LALMs) along three complementary dimensions: (1) Instruction following: Text Instruction Following (TIF), Speech Instruction Following (SIF), English Text Instruction Following (TIF-En)… See the full description on the dataset page:

Source: Hugging Face Hub (tencent/VCB-Bench). Metadata imported from the dataset’s Hub tags.