Balrog
Harbor: pendingTracked by Epoch AI's Benchmarking Hub (results collected from external sources).
Incomplete29%Knowledge & Reasoning
- Samples
- N/A
- Released
- 2024
- Scores
- 39
Vinoground
A temporal counterfactual dataset composing of 1000 short and natural video-caption pairs.
How to run
ltb · vinoground
$ npm i -g lettokenburn-cli$ ltb bench info vinoground
External harness. No Harbor adapter has been generated for this benchmark yet. `ltb bench info` prints the known links.
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
Scores
16 sourced entries| Harness | Source | Confidence | Link | |||
|---|---|---|---|---|---|---|
| GPT-4o | 59.2 | N/A | Independent | medium | N/A | paperswithcode.com |
| GPT-4o | 54.0 | N/A | Independent | medium | N/A | paperswithcode.com |
| Qwen2-VL-72B | 50.4 | N/A | Independent | medium | Sep 18, 2024 | arxiv.org |
| LLaVA-OneVision-Qwen2-72B | 48.4 | N/A | Independent | medium | Aug 6, 2024 | arxiv.org |
| LLaVA-OneVision-Qwen2-7B | 41.6 | N/A | Independent | medium | Aug 6, 2024 | arxiv.org |
| Qwen2-VL-7B | 40.2 | N/A | Independent | medium | Sep 18, 2024 | arxiv.org |
| Gemini 1.5 Pro | 37.0 | N/A | Independent | medium | Mar 8, 2024 | arxiv.org |
| Gemini 1.5 Pro | 35.8 | N/A | Independent | medium | Mar 8, 2024 | arxiv.org |
| Claude 3.5 Sonnet | 32.8 | N/A | Independent | medium | N/A | paperswithcode.com |
| InternLM-XC-2.5 | 30.8 | N/A | Independent | medium | Jul 3, 2024 | arxiv.org |
| InternLM-XC-2.5 | 28.8 | N/A | Independent | medium | Jul 3, 2024 | arxiv.org |
| LLaVA-NeXT-Video-34B | 25.8 | N/A | Independent | medium | N/A | paperswithcode.com |
| Phi-3.5-Vision | 24.0 | N/A | Independent | medium | N/A | paperswithcode.com |
| LLaVA-NeXT-Video-34B | 23.0 | N/A | Independent | medium | N/A | paperswithcode.com |
| LLaVA-NeXT-Video-7B | 21.8 | N/A | Independent | medium | N/A | paperswithcode.com |
| LLaVA-NeXT-Video-7B | 21.8 | N/A | Independent | medium | N/A | paperswithcode.com |
Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.
Reproduced by the community
All runsLoading community runs…
About
A temporal counterfactual dataset composing of 1000 short and natural video-caption pairs.