ddz16/DyBench
DyBench Project Page | Paper | GitHub DyBench is a paired counterfactual video benchmark introduced in the paper "Learning Spatiotemporal Sensitivity in Video LLMs via Counterfactual Reinforcement Learning". The benchmark is designed to evaluate the spatiotemporal sensitivity of Video Large Language Models (Video LLMs). It addresses the issue of models relying on "shortcuts" (such as single-frame cues or language priors) rather than tracking actual video dynamics. DyBench… See the full description on the dataset page: https://huggingface.co/datasets/ddz16/DyBench.
| Author / organization | ddz16 |
|---|---|
| License (as declared) | mit |
| Downloads | 147 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 26 October 2025 |
| Last updated at source | 25 May 2026 |
Recent changes
- 16 Oct 2026 Description: DyBench Project Page | Paper | GitHub DyBench is a paired counterfactual video benchmark introduced in the paper "Learning Spatiotemporal Sensitivity in Video LLMs via Counterfactual Reinforcement Learning". The benchmark is designed to eva → DyBench Project Page | Paper | GitHub DyBench is a paired counterfactual video benchmark introduced in the paper "Learning Spatiotemporal Sensitivity in Video LLMs via Counterfactual Reinforcement Learning". The benchmark is designed to eva
Changes detected by Silytics between refreshes of the source.
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 16 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.