Livibench: Benchmark Data for Interactive Livestream Video Understanding

Livibench is benchmark data used in the paper 'LiViBench: An Omnimodal Benchmark for Interactive Livestream Video Understanding'. The dataset was authored by Xiaodong Wang and others, with a repository last updated on February 2, 2026.

Use Cases

• Benchmarking video understanding models based on the omnimodal livestream data.
• Training models for interactive livestream analysis based on the described benchmark tasks.
• Researching multimodal fusion techniques for video, audio, and text in a livestream context.

Strengths

• Dataset is directly associated with a named, cited academic paper (arXiv:2601.15016).
• Repository update date is precisely recorded as 2026-02-02 08:57:05.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.

Provenance

Source

Paper: LiViBench: An Omnimodal Benchmark for Interactive Livestream Video Understanding (arXiv:2601.15016).

Collection Method

Likely collected and annotated for research benchmarking purposes.

Time Range

null

Freshness

Last updated 2026-02-02 08:57:05; freshness should be verified.

Geography

null