📊
VoxEval dataset by KeyNG101
⭐ 34
📊
Dataset

VoxEval

by KeyNG101 hf-dataset--keyng101--voxeval
Nexus Index
34.0 Top 100%
S: Semantic 50
A: Authority 0
P: Popularity 62
R: Recency 53
Q: Quality 30
Tech Context
Vital Performance
0 DL / 30D
0.0%
Source →
Data Integrity 34 FNI Score
- Size
- Rows
Parquet Format
- Tokens
Dataset Information Summary
Entity Passport
Registry ID hf-dataset--keyng101--voxeval
License CC-BY-4.0
Provider huggingface
📜

Cite this dataset

Academic & Research Attribution

BibTeX 
@misc{hf_dataset__keyng101__voxeval,
  author = {KeyNG101},
  title = {VoxEval Dataset},
  year = {2026},
  howpublished = {\url{https://huggingface.co/datasets/keyng101/voxeval}},
  note = {Accessed via Free2AITools Knowledge Fortress}
}
APA Style
KeyNG101. (2026). VoxEval [Dataset]. Free2AITools. https://huggingface.co/datasets/keyng101/voxeval

đŸ”ŦTechnical Deep Dive

Full Specifications [+]

⚖ī¸ Nexus Index V2.0

Methodology Index Protocol
34.0
TOP 100% SYSTEM IMPACT
Semantic (S) 50
Authority (A) 0
Popularity (P) 62
Recency (R) 53
Quality (Q) 30

đŸ’Ŧ Index Insight

FNI V2.0 for VoxEval: Semantic (S:50), Authority (A:0), Popularity (P:62), Recency (R:53), Quality (Q:30).

Free2AITools Nexus Index

Verification Authority

HuggingFace API GitHub Metadata Arxiv Citation DB System Audit
Unbiased Data Node Refresh: VFS Live
âŦ‡ī¸
Downloads
214,022

👁ī¸ Data Preview

📊

Row-level preview not available for this dataset.

Schema structure is shown in the Field Logic panel when available.

🔗 Explore Full Dataset ↗

đŸ§Ŧ Field Logic

đŸ§Ŧ

Schema not yet indexed for this dataset.

Dataset Specification

VoxEval

GitHub arXiv

Github repository for paper: VoxEval: Benchmarking the Knowledge Understanding Capabilities of End-to-End Spoken Language Models

Also check out our survey paper at Recent Advances in Speech Language Models: A Survey!

VoxEval is a novel speech question-answering benchmark specifically designed to assess SLMs' knowledge understanding through purely speech-based interactions.

Below are the three highlights of our VoxEval benchmark:

  • End-to-end speech-based evaluation: Both input and output are audio-based.
  • Diverse audio conditions: VoxEval includes audio files featuring a variety of speakers, speaking styles, and audio qualities.
  • Complex spoken evaluations: It supports advanced spoken assessments, including spoken math tasks.

Download Data

You can access our VoxEval Dataset Repository on 🤗 Hugging Face to directly download the dataset.

Below is the layout of the dataset folder. all_fewshot_examples folder contains the few shot examples for the evaluation. math_CoT_fewshot folder contains the few shot examples for evaluating the math subjects via Chain-of-Thought prompting. test folder contains the actual test data in VoxEval.

text 
  # ├── Root Folder
  # │   ├── all_fewshot_examples
  # │       ├── alloy (different speaker voice)
  # │           ├── abstract_algebra_4o (different subjects)
  # │               ├── XXX.mp3
  # │               ├── ...
  # │           ├── ...
  # │       ├── echo
  # │       ├── ...
  # │   ├── test
  # │       ├── alloy (different settings)
  # │           ├── abstract_algebra_4o (different subjects)
  # │               ├── XXX.mp3
  # │               ├── ...
  # │           ├── ...
  # │       ├── echo
  # │       ├── ...

Evaluation results of existing end-to-end Spoken Language Models

SLMs SpeechGPT TWIST SPIRIT-LM Moshi GLM-4-Voice
Speakers
Alloy 0.0001 0.0480 0.2084 0.1216 0.3763
Echo 0.0001 0.0558 0.2096 0.1221 0.3764
Fable 0.0000 0.0116 0.2084 0.1153 0.3642
Nova 0.0001 0.0332 0.2070 0.1298 0.3677
Onyx 0.0002 0.0275 0.1966 0.1192 0.3764
Shimmer 0.0000 0.0516 0.2076 0.1205 0.3815
Speaking Styles
Linguistic 0.0001 0.0488 0.2044 0.1187 0.3643
Speed 0.0001 0.0503 0.1911 0.1013 0.3469
Pitch 0.0000 0.0544 0.1788 0.0609 0.3345
Audio Qualities
Noise 0.0000 0.0368 0.1950 0.1018 0.3695
Other Env Acoustics 0.0001 0.0434 0.2019 0.1051 0.3728
Underlying Text LMs Llama-7B Llama-7B Llama-2-7B Helium-7B GLM-4-9B
Text MMLU 0.3510 0.3510 0.4530 0.5430 0.7470

What's Next

  • The VoxEval evaluation code will be released soon.

License

The dataset is licensed under the Creative Commons Attribution 4.0.

Citation

text 
@article{cui2025voxeval,
  title={VoxEval: Benchmarking the Knowledge Understanding Capabilities of End-to-End Spoken Language Models},
  author={Cui, Wenqian and Jiao, Xiaoqi and Meng, Ziqiao and King, Irwin},
  journal={arXiv preprint arXiv:2501.04962},
  year={2025}
}

Social Proof

HuggingFace Hub
214.0KDownloads
Hub Discussions
🤗 Data Source: Hugging Face ↗
🔄 Daily sync (03:00 UTC)

AI Summary: Based on Hugging Face metadata. Not a recommendation.

📊 FNI Methodology 📚 Knowledge Baseℹī¸ Verify with original source

🛡ī¸ Dataset Transparency Report

Technical metadata sourced from upstream repositories.

Open Metadata

🆔 Identity & Source

id
hf-dataset--keyng101--voxeval
slug
keyng101--voxeval
source
huggingface
author
KeyNG101
license
CC-BY-4.0
tags
license:cc-by-4.0, arxiv:2501.04962, arxiv:2410.03751, region:us

⚙ī¸ Technical Specs

architecture
null
params billions
null
context length
null
pipeline tag

📊 Engagement & Metrics

downloads
214,022
stars
0
forks
0

Data indexed from public sources. Updated daily.