# Referring Description on Screenshot Out-of-Domain Source: [Referring Description on Screenshot Out-of-Domain]() Collection index: [llms.txt]() Data updated: 2026-02-27T14:24:07.585Z Evaluation of Referring Description generation performance on the Screenshot Out-of-Domain dataset, measured using BLEU-4, METEOR, CIDEr, and SPICE. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Referring Description | | Dataset (including any stated split) | Screenshot Out-of-Domain | | Metric | BLEU-4 | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 9 of 9 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [BLEU-4](): Higher is better; unit: PERCENT. - [METEOR](): Higher is better; unit: SCALAR. - [CIDEr](): Higher is better; unit: SCALAR. - [SPICE](): Higher is better; unit: SCALAR. ## Recorded results | Method | BLEU-4 | Result date | Report paper | | --- | --- | --- | --- | | Qwen2.5-VL + ControlMLLM++ | 9.05 | Not recorded | 699d30c56017c4460ffab7ec | | Qwen2.5-VL + ControlMLLM | 4.26 | Not recorded | 699d30c56017c4460ffab7ec | | Qwen2.5-VL | 1.09 | Not recorded | 699d30c56017c4460ffab7ec | | LLaVA-HR + ControlMLLM++ | 0.98 | Not recorded | 699d30c56017c4460ffab7ec | | LLaVA-HR + ControlMLLM | 0.64 | Not recorded | 699d30c56017c4460ffab7ec | | LLaVA-1.5 + ControlMLLM++ | 0.57 | Not recorded | 699d30c56017c4460ffab7ec | | LLaVA-1.5 + ControlMLLM | 0.45 | Not recorded | 699d30c56017c4460ffab7ec | | LLaVA-1.5 | 0.32 | Not recorded | 699d30c56017c4460ffab7ec | | LLaVA-HR | 0.29 | Not recorded | 699d30c56017c4460ffab7ec | ### Qwen2.5-VL + ControlMLLM++ BLEU-4: 9.05. Result ID: 69a1a90535dfcabb00e810a3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen2.5-VL","Enhanced with":"ControlMLLM++"}. This record does not assert independent reproduction. ### Qwen2.5-VL + ControlMLLM BLEU-4: 4.26. Result ID: 69a1a90535dfcabb00e810a2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen2.5-VL","Enhanced with":"ControlMLLM"}. This record does not assert independent reproduction. ### Qwen2.5-VL BLEU-4: 1.09. Result ID: 69a1a90535dfcabb00e810a1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen2.5-VL"}. This record does not assert independent reproduction. ### LLaVA-HR + ControlMLLM++ BLEU-4: 0.98. Result ID: 69a1a90535dfcabb00e810a0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"LLaVA-HR","Enhanced with":"ControlMLLM++"}. This record does not assert independent reproduction. ### LLaVA-HR + ControlMLLM BLEU-4: 0.64. Result ID: 69a1a90535dfcabb00e8109f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"LLaVA-HR","Enhanced with":"ControlMLLM"}. This record does not assert independent reproduction. ### LLaVA-1.5 + ControlMLLM++ BLEU-4: 0.57. Result ID: 69a1a90535dfcabb00e8109d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"LLaVA-1.5","Enhanced with":"ControlMLLM++"}. This record does not assert independent reproduction. ### LLaVA-1.5 + ControlMLLM BLEU-4: 0.45. Result ID: 69a1a90535dfcabb00e8109c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"LLaVA-1.5","Enhanced with":"ControlMLLM"}. This record does not assert independent reproduction. ### LLaVA-1.5 BLEU-4: 0.32. Result ID: 69a1a90535dfcabb00e8109b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Base Model":"LLaVA-1.5"}. This record does not assert independent reproduction. ### LLaVA-HR BLEU-4: 0.29. Result ID: 69a1a90535dfcabb00e8109e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Base Model":"LLaVA-HR"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.