# Multimodal Concept Understanding on MCUB 3 Source: [Multimodal Concept Understanding on MCUB 3]() Collection index: [llms.txt]() Data updated: 2026-03-24T15:47:31.078Z Evaluation of model performance on the MCUB 3 dataset for Multimodal Concept Understanding, measured by component scores including I+V+P+L, A+V+P+L, I+A+P+L, I+A+V+L, and Average. 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 | Multimodal Concept Understanding | | Dataset (including any stated split) | MCUB 3 | | Metric | I+V+P+L Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 10 of 10 | | 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 - [I+V+P+L Score](): Higher is better; unit: PERCENT. - [A+V+P+L Score](): Higher is better; unit: PERCENT. - [I+A+P+L Score](): Higher is better; unit: PERCENT. - [I+A+V+L Score](): Higher is better; unit: PERCENT. - [Average Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | I+V+P+L Score | Result date | Report paper | | --- | --- | --- | --- | | SSAM | 59.4 | Not recorded | 69c21aa10636e5bf4afc49fd | | DAMC | 58.2 | Not recorded | 69c21aa10636e5bf4afc49fd | | NaiveMC | 53 | Not recorded | 69c21aa10636e5bf4afc49fd | | OptMerge | 52.8 | Not recorded | 69c21aa10636e5bf4afc49fd | | TSV | 48 | Not recorded | 69c21aa10636e5bf4afc49fd | | Proj-Only | 42.6 | Not recorded | 69c21aa10636e5bf4afc49fd | | WUDI | 38.2 | Not recorded | 69c21aa10636e5bf4afc49fd | | Task Arithmetic | 36.8 | Not recorded | 69c21aa10636e5bf4afc49fd | | ImageBind-LLM | 31.8 | Not recorded | 69c21aa10636e5bf4afc49fd | | X-InstructBLIP | 29.4 | Not recorded | 69c21aa10636e5bf4afc49fd | ### SSAM I+V+P+L Score: 59.4. Result ID: 69c2b20047f34ce924abe23e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Type":"Training-free model merging methods"}. This record does not assert independent reproduction. ### DAMC I+V+P+L Score: 58.2. Result ID: 69c2b20047f34ce924abe23d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Type":"Training-free model merging methods"}. This record does not assert independent reproduction. ### NaiveMC I+V+P+L Score: 53. Result ID: 69c2b20047f34ce924abe23c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Type":"Training-free model merging methods"}. This record does not assert independent reproduction. ### OptMerge I+V+P+L Score: 52.8. Result ID: 69c2b20047f34ce924abe23b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Type":"Training-free model merging methods"}. This record does not assert independent reproduction. ### TSV I+V+P+L Score: 48. Result ID: 69c2b20047f34ce924abe239. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Type":"Training-free model merging methods"}. This record does not assert independent reproduction. ### Proj-Only I+V+P+L Score: 42.6. Result ID: 69c2b20047f34ce924abe237. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Type":"Models finetuned on paired multimodal data"}. This record does not assert independent reproduction. ### WUDI I+V+P+L Score: 38.2. Result ID: 69c2b20047f34ce924abe23a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Type":"Training-free model merging methods"}. This record does not assert independent reproduction. ### Task Arithmetic I+V+P+L Score: 36.8. Result ID: 69c2b20047f34ce924abe238. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Type":"Training-free model merging methods"}. This record does not assert independent reproduction. ### ImageBind-LLM I+V+P+L Score: 31.8. Result ID: 69c2b20047f34ce924abe235. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Type":"Models finetuned on paired multimodal data"}. This record does not assert independent reproduction. ### X-InstructBLIP I+V+P+L Score: 29.4. Result ID: 69c2b20047f34ce924abe236. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Type":"Models finetuned on paired multimodal data"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.