# Offense Classification on MMSD Offense style 2.0 (test) Source: [Offense Classification on MMSD Offense style 2.0 (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:30:58.169Z Evaluation of Offense Classification performance on the test split of the MMSD Offense style 2.0 dataset, reporting Accuracy and F1 scores. 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 | Offense Classification | | Dataset (including any stated split) | MMSD Offense style 2.0 (test) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 8 of 8 | | 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 - [Accuracy](): Higher is better; unit: PERCENT. - [F1 Score](): Higher is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | SFT on Combined → GRPO (Combined) | 61.49 | Not recorded | 697846b1a2b28e761d87c5f4 | | Gemini 2.5 Flash | 60.95 | Not recorded | 697846b1a2b28e761d87c5f4 | | SFT on Combined → GRPO (Style specific) | 60.27 | Not recorded | 697846b1a2b28e761d87c5f4 | | LLaMA 90B Vision Instruct | 55.47 | Not recorded | 697846b1a2b28e761d87c5f4 | | Qwen2.5-VL 32B Instruct | 49.37 | Not recorded | 697846b1a2b28e761d87c5f4 | | Phi-4 Multimodal | 48.74 | Not recorded | 697846b1a2b28e761d87c5f4 | | Gemma 3 27B | 45.33 | Not recorded | 697846b1a2b28e761d87c5f4 | | LLaVA 1.5 7B | 43.22 | Not recorded | 697846b1a2b28e761d87c5f4 | ### SFT on Combined → GRPO (Combined) Accuracy: 61.49. Result ID: 699670555698665ec4b766f7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompting":"zero-shot","Setup":"Combined"}. This record does not assert independent reproduction. ### Gemini 2.5 Flash Accuracy: 60.95. Result ID: 699670555698665ec4b766f0. [Reporting paper]() Recorded evaluation configuration: {"Prompting":"zero-shot"}. This record does not assert independent reproduction. ### SFT on Combined → GRPO (Style specific) Accuracy: 60.27. Result ID: 699670555698665ec4b766f6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompting":"zero-shot","Setup":"Style specific"}. This record does not assert independent reproduction. ### LLaMA 90B Vision Instruct Accuracy: 55.47. Result ID: 699670555698665ec4b766f2. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Prompting":"zero-shot"}. This record does not assert independent reproduction. ### Qwen2.5-VL 32B Instruct Accuracy: 49.37. Result ID: 699670555698665ec4b766f1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompting":"zero-shot"}. This record does not assert independent reproduction. ### Phi-4 Multimodal Accuracy: 48.74. Result ID: 699670555698665ec4b766f4. [Reporting paper]() Recorded evaluation configuration: {"Prompting":"zero-shot"}. This record does not assert independent reproduction. ### Gemma 3 27B Accuracy: 45.33. Result ID: 699670555698665ec4b766f3. [Reporting paper]() Recorded evaluation configuration: {"Prompting":"zero-shot"}. This record does not assert independent reproduction. ### LLaVA 1.5 7B Accuracy: 43.22. Result ID: 699670555698665ec4b766f5. [Reporting paper]() Recorded evaluation configuration: {"Prompting":"zero-shot"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.